Integrating Complexity and Risk Analysis for Selection of Management Approaches in Complex Projects: Application to UN Peacekeeping Missions

Abstract

The growing complexity and dynamism of industrial and organizational projects require management approaches that can effectively adapt to uncertainty and rapidly changing operational environments. In this context, this study proposes a methodology to identify the most suitable management approach—predictive, agile, or hybrid—in complex projects. Building on the “Approach suitability tool” of the Project Management Institute’s (PMI), the methodology integrates quantitative assessments of complexity and systemic risk. This is achieved through the analysis of stakeholder and risk networks, using metrics such as cyclomatic complexity and the coevolution parameter g, which allow for a deeper understanding of interactions and the evolution of project elements. The methodology was validated in three peacekeeping missions of the United Nations: UNMISS in South Sudan, MONUSCO in the Democratic Republic of Congo, and MINUSTAH in Haiti. The results confirm that the methodology accurately identifies the most appropriate management approach, emphasizing the effectiveness of hybrid approaches in complex and volatile environments. The proposed methodology serves as a valuable tool for optimizing project management in diverse contexts, enabling a quantitative and systematic evaluation of complexity and risk. It is adaptable and applicable to a wide range of complex projects, improving decision-making and planning in uncertain settings. Furthermore, by incorporating resilience as a cross-cutting principle, the methodology strengthens the ability of projects and their teams to maintain functionality and sustain learning even in highly volatile environments, where continuous adaptation becomes a critical success factor. In this sense, resilience emerges as the property that allows projects to absorb disruptions, reorganize, and preserve their core purpose without losing cohesion or direction.

1. Introduction

The project management literature has shown that, although projects share a common basic structure, they may differ significantly in their level of complexity and in the nature of the interactions that characterize them. These differences are not trivial, as they directly condition the suitability of the management approaches applied in each case. In this context, project complexity has become a key concept, clearly multidimensional in nature, extending beyond size, duration or technical difficulty. Complexity is primarily associated with the degree of interdependence between project elements, the non-linearity of their interactions, their dynamic character, and the coexistence of multiple actors with heterogeneous interests and capabilities [1,2,3]. From this perspective, complexity describes a set of structural and dynamic conditions that limit the ability to anticipate project behavior using simple deterministic models and that require management approaches capable of operating under uncertainty and change.

Risk, in turn, constitutes an analytical concept distinct from the project itself and, in all cases, is defined by the uncertainty associated with the occurrence of certain events and their impact on the established objectives. This definition applies equally to conventional projects and to complex projects. The difference between these contexts, therefore, does not lie in the existence of risk or in its potential scope, but in the way risk behaves over the project life cycle. In projects with relatively stable structures and low interdependence, risks can be identified as discrete events whose essential characteristics—origin, probability and impact—remain essentially unchanged, regardless of whether they ultimately materialize or not. In complex projects, by contrast, risk may evolve, combine with other risks, or appear in an unanticipated manner because of the internal dynamics of the project system itself [4,5]. The high level of interdependence between components, together with the continuous evolution of relationships among actors, means that risk exhibits a dynamic and non-stationary behavior, in which its identification and characterization cannot be considered definitive from the early stages.

Traditionally, project management has relied on approaches grounded in predictability and control, such as the waterfall method [6], the Critical Path Method (CPM) [7], and the Program Evaluation and Review Technique (PERT) [8], which are effective in contexts where objectives and resources remain relatively stable [9]. However, the nature of projects has evolved significantly, particularly in technological sectors and in environments characterized by high volatility, introducing levels of uncertainty that challenge the effectiveness of conventional methods [10]. In complex projects, uncertainty makes it difficult to identify all risks exhaustively from the early stages, which requires methodologies capable of managing both structural and dynamic complexity [11]. Traditional predictive approaches, based on rigid requirements defined at the beginning of the project life cycle, often show limitations when applied to such contexts [12]. In response, alternative approaches—iterative, incremental and agile—have emerged, facilitating adaptation to change and the progressive reduction in risk [13,14]. Hybrid approaches, which combine elements of traditional and adaptive methods, have become particularly relevant for projects that simultaneously require structure and flexibility [15].

In this context, adaptability and resilience become fundamental characteristics of modern project management. Resilience refers to the capacity of the team and the project system to absorb disturbances and maintain progress towards objectives despite adversity [16], while adaptability involves the continuous adjustment of processes, resources and goals in response to changes in the environment [17]. Resilience should not be understood merely as resistance or as a return to a previous state, but as the capacity of the project to absorb disturbances, reorganize, and continue generating value under new conditions. From a complexity perspective, resilience emerges as an adaptive property of the system, enabling learning and reconfiguration in response to both anticipated and unforeseen risks.

Success in complex projects cannot be assessed solely in terms of initial compliance with time and budget constraints but rather depends on the achievement of fundamental objectives, even when these must be reassessed during execution [18]. This implies a shift in focus towards value creation and the satisfaction of stakeholders’ dynamic needs [19]. In this regard, the analysis of complexity is essential, as the early identification of the factors that generate it enables the design of more effective strategies to mitigate risks and improve project performance [1]. Tools such as complexity classifications, uncertainty matrices and risk mapping play a relevant role in this process [16].

Network analysis provides a particularly suitable framework for identifying interdependencies and relationships among project components and actors, making it possible to reveal patterns that may significantly influence project evolution [20]. Such approaches are especially useful in complex projects, where emergent effects and cascading phenomena arise from dense interaction structures rather than from isolated elements. However, their application requires rigorous expertise and additional resources, which may hinder implementation in certain contexts [21]. In highly uncertain environments, effective communication and collaboration among teams and stakeholders play a critical role. Multidisciplinary coordination and collaborative approaches foster transparency and joint problem-solving [22], while digital tools and collaboration platforms act as key enablers [23]. Within this framework, technology plays a relevant role in the management of complex projects.

Furthermore, risk management occupies a central position in complex project environments, as it constitutes the primary instrument for anticipating and mitigating the uncertainties inherent in these systems. Unlike simpler projects, where risks are generally known and manageable through traditional techniques, in complex projects risks may interact, amplify one another, and give rise to unforeseen situations. Understanding this dynamic nature of risk and having tools that allow its continuous identification, assessment and management—considering technical, organizational and strategic dimensions—is therefore essential. From this perspective, risk management and resilience can be understood as complementary dimensions: while the former seeks to anticipate and reduce impacts, the latter enables the project to sustain performance when disturbances occur and to transform experience into organizational learning.

From this standpoint, the selection of an appropriate project management approach becomes a critical decision problem. Although several conceptual frameworks are widely used to address complexity and uncertainty, such as the Cynefin framework [24] or the Stacey Matrix [25], these models primarily act as classificatory or sense-making tools and do not provide an operational integration of complexity and risk. Other contributions, such as the work of Boehm and Turner [26], advance the relationship between project characteristics and management approaches, although their applicability is largely limited to software-intensive contexts. Similarly, Appendix X3 of the PMI Agile Practice Guide [27] introduces an approach suitability assessment but does not explicitly integrate complexity and risk as central decision variables. Overall, these approaches rely predominantly on qualitative or semi-qualitative judgements, which limits reproducibility and comparability across projects, and treat complexity and risk as contextual descriptors rather than as measurable systemic properties derived from project structure and dynamics.

To address these limitations, this work develops a methodology that extends the scope of the approach suitability tool proposed in the PMI Agile Practice Guide by incorporating complexity and risk as fundamental categories for selecting the management approach—predictive, agile or hybrid. The methodology operationalizes these concepts through network-based metrics, establishing a transparent and reproducible linkage between the structural and dynamic characteristics of the project and the selection of the most appropriate management approach. For validation purposes, the proposal is applied to three United Nations peacekeeping missions (UNMISS, MONUSCO and MINUSTAH), characterized by high levels of complexity, multiple actors and highly unstable contexts. The methodology aims to facilitate more informed decision-making in dynamic environments and provides a framework applicable to other highly complex projects.

The remainder of this article is structured as follows. Section 2 establishes the theoretical and analytical framework related to project complexity and risk. Section 3 presents the proposed methodology. Section 4 validates the proposal through the case studies. Section 5 discusses the results, highlighting the advantages and limitations of the methodology, and Section 6 presents the main conclusions.

2. Materials and Methods

The development of this research followed the methodological workflow shown in Figure 1, structured into four main phases aimed at designing, substantiating, and validating a methodology capable of selecting the most appropriate management approach for complex projects.

In PHASE 1, an exhaustive review of the scientific and professional literature on complex project management was conducted, with particular attention to models for selecting a management approach (predictive, agile, and hybrid) and tools that incorporate complexity and risk analysis. This review made it possible to identify the limitations of existing methods—such as the Stacey Matrix, the Cynefin framework, and the PMI approach suitability tool—and to establish the conceptual foundations of the new proposal.

Based on the conclusions of the scientific and technical literature review, PHASE 2 involved selecting the relevant variables for evaluating project complexity and systemic risk. In this phase, the quantitative metrics—cyclomatic complexity, the coevolution parameter g, structural risk, and dynamic risk—were defined, along with the criteria for their integration into the PMI tool. The objective was to translate complex systems theory into operational parameters applicable to the project management context. Implicitly, this formulation incorporates resilience as an emergent property of the system, since the interaction between complexity and risk determines the project’s ability to maintain functional stability under disturbances and to reconfigure itself without losing its overall purpose.

PHASE 3 redesigns the Project Management Institute’s Approach Suitability Tool by replacing two of its original parameters (“change” and “criticality”) with the new variables “total complexity” and “total risk.” A calculation procedure consistent with the internal logic of the tool was developed, ensuring compatibility with its evaluation scale.

Finally, in PHASE 4, the methodology was validated through its application to three case studies: the United Nations peacekeeping missions UNMISS, MONUSCO, and MINUSTAH. These missions were selected due to their high levels of complexity and contextual diversity, which enabled assessment of the proposal’s effectiveness and adaptability. Validation included the analysis of stakeholder and risk networks using UCINET software (version 6, Analytic Technologies, Harvard, MA, USA), as well as comparison between the tool’s results and the management approaches applied in each mission.

3. Results

The implementation of the methodological phases described in the previous section yielded a set of results organized into three main blocks. First, Section 3.1 presents a state-of-the-art review that compiles and analyses the main existing methods for selecting management approaches in complex projects, identifying their strengths and limitations. Second, Section 3.2 presents the proposed methodological framework, based on the integration of complexity and risk metrics into the Project Management Institute (PMI) Approach Suitability Tool. Finally, Section 4 describes the empirical validation carried out through the application of the proposed method to three United Nations peacekeeping missions (UNMISS, MONUSCO, and MINUSTAH), which made it possible to assess its coherence and effectiveness in contexts of high uncertainty and volatility.

3.1. Foundations of Complexity Measurement in Projects

Project complexity has become established in the literature as a systemic and multidimensional concept that cannot be explained solely in terms of traditional variables such as project size, duration, or technical difficulty. In certain projects, overall system behavior does not result from the simple aggregation of its parts, but rather from non-linear interactions among multiple interdependent elements (including actors, tasks, resources, and environmental conditions) whose relationships evolve over time. This interaction structure gives rise to adaptive dynamics, feedback processes, and emergent phenomena that limit the ability to anticipate project behavior through simple cause–effect relationships [28,29].

From this perspective, complexity should not be understood as a static attribute or as a mere accumulation of difficulties, but as the result of structural and dynamic configurations that condition the evolution of the project system. This implies a close, though not equivalent, relationship between complexity and uncertainty: while uncertainty is associated with a lack of information about future states, complexity reflects the intrinsic difficulty of anticipating system behavior even when such information is available, due to interdependence, non-linearity, and ongoing adaptive processes [30,31]. These characteristics challenge traditional project management approaches based on decomposition and control and justify the need for analytical frameworks capable of capturing both the structural and dynamic dimensions of projects.

Projects are not homogeneous. Their type and characteristics largely determine which management approaches are most suitable. Some projects are less complex, where the factors affecting outcomes are relatively predictable, activities are well defined, and traditional planning and control methods are generally sufficient [32]. At the other extreme are projects developed in dynamic and highly uncertain environments, characterized by large scale, the participation of multiple stakeholders, and a high degree of interdependence between activities [33]. In such cases, conventional methods present clear limitations, and an adaptive and holistic approach is required to cope with the inherent challenges of these types of projects [34].

External factors, such as rapid technological advances or changes in the regulatory framework, further increase complexity in certain projects and complicate the application of traditional methods [35]. As interdependence grows, a problem or unexpected event localized in one part of the project may trigger cascading effects in other areas, significantly amplifying its impact [36]. This kind of interaction between elements is precisely what distinguishes complex projects, where it is not enough to address problems in isolation; instead, it is necessary to understand interrelationships and the overall behavior of the system [37].

This reasoning justifies the need, in certain projects, for additional mechanisms to anticipate and manage emerging situations. Anticipation and adaptability become especially relevant in highly complex environments, where conditions change rapidly and unforeseen situations arise that cannot be easily managed using traditional approaches [38]. The early identification and continuous monitoring of critical factors throughout the project life cycle are therefore fundamental [39].

The way project complexity manifests itself varies across application domains, and these differences have direct implications for both its effects and the suitability of management approaches. In engineering and construction projects, complexity is primarily associated with structural interdependencies, contractual interfaces, and coordination among multiple organizations, which tends to favor structured approaches or hybrid configurations. In new product development projects, complexity is more closely linked to technological uncertainty, evolving requirements, and innovation dynamics, increasing the relevance of iterative and hybrid management strategies. Information systems and information technology projects are typically characterized by strong socio-technical interactions, frequent changes, and stakeholder-driven uncertainty, making agile or hybrid approaches particularly appropriate. By contrast, projects operating in unstable and rapidly evolving environments are dominated by dynamic and temporal complexity, where decision-making under high uncertainty and time pressure requires highly adaptive and flexible management approaches. Despite these domain-specific differences, complex projects across sectors share a set of common characteristics—such as interdependence, non-linearity, and emergent behavior—that justify the use of generalizable analytical frameworks for complexity and risk assessment.

Measuring complexity in projects is fundamental for effective management and the achievement of expected outcomes, since complexity can significantly affect coordination, communication, and adaptability [40]. Projects can be represented as complex networks, specifically small-world networks, characterized by high levels of local clustering and short path lengths between nodes [41]. These properties reflect the intricate interconnections and interactions among project elements and participants. UCINET software [42] allows for the structural analysis of complex networks through visualization and the calculation of specific metrics, facilitating a more accurate interpretation of the project’s relational architecture.

Project complexity is addressed in this study from a dimensional perspective, in which total project complexity is understood as the sum of two complementary components: structural complexity and dynamic complexity. The former is associated with the configuration of interdependencies within the project system, while the latter reflects its evolutionary behavior over time.

Structural complexity is characterized through network-based metrics, which represent the project as a set of interconnected actors, units, and functions. In this context, ECyM is used as an indicator of structural complexity, as it captures the degree of branching and interdependence of the project network. As network density increases, the number of branching points grows, or strongly connected components become more prevalent, structural complexity increases accordingly, reflecting a higher coordination and control burden.

Dynamic complexity is addressed through the co-evolution parameter g, which represents the degree of coupling between changes in the state of the project and adaptations in its underlying structure. Higher values of g indicate greater dynamic complexity, as they reflect more intense and continuous adaptation processes during project execution. The combination of both dimensions allows total project complexity to be assessed in a coherent and operational manner, and is particularly suitable for the analysis of complex, multi-actor projects such as United Nations peacekeeping missions.

To measure structural complexity, this study uses extended cyclomatic complexity (ECyM), which quantifies network complexity based on its topology [43,44]. ECyM is calculated using Expression (1):

$$\mathrm{E}\mathrm{C}\mathrm{y}\mathrm{M}\left(\mathrm{P}\mathrm{N}\right)=\mathrm{E}-\mathrm{N}+\mathrm{p},$$

(1)

where E is the number of edges in the network, N is the number of nodes, and p is the number of cliques or strongly connected components (SCC). This last measure reflects the degree of interconnectivity and the presence of cycles in the network, indicating the inherent complexity of coordination and information flows within the project. A higher ECyM value suggests greater structural complexity, which can pose challenges for project management due to increased interdependencies.

However, considering structural complexity alone is not sufficient, since network dynamics also play a crucial role in overall project complexity. Coevolution captures dynamic complexity by considering how the network structure and the state of its nodes evolve over time [45]. The parameter g, which represents coevolution in a network, is defined in Expression (2):

$$\mathrm{g} = {\displaystyle \frac{〈\mathrm{k}〉}{\left[\mathrm{N}〈\mathrm{C}〉\right]}},$$

(2)

where 〈k〉 range is the average degree of the network—which, in small-world networks, can be considered a higher-order degree—N is the number of nodes, and 〈C〉 range is the average clustering coefficient. In small-world networks with a minimum clustering value of around 0.75, it can be observed that as N increases, g slowly tends to 0, as shown in Figure 2.

The figure shows that, in networks with a small number of nodes, the parameter g takes relatively high values and exhibits high variability, indicating intense coevolution associated with the limited size of the network. As the number of nodes increases, the value of g decreases and progressively converges towards more stable ranges, suggesting that coevolutionary processes become less sensitive to local perturbations. This evolution reflects a gradual reduction in dynamic complexity associated with the initial growth phase of the network.

Regarding existing methodologies designed to guide the selection of project management approaches, several frameworks have contributed to classifying project contexts according to levels of uncertainty and complexity. Models such as the Cynefin framework [24] or the Stacey Matrix [25] offer valuable sense-making and classificatory perspectives, helping practitioners interpret project environments and recognize when adaptive approaches may be required. However, these models do not provide an operational assessment of structural and dynamic complexity, nor do they explicitly integrate systemic risk, which limits their applicability as analytical foundations for quantitative decision-support tools in highly complex project environments.

Regarding existing methodologies designed to guide the selection of the most suitable project management approach based on the characteristics of the organizational context and the project itself, various authors have proposed different strategies. Ranging from conceptual models to graphical decision tools, these methods converge on the need to adapt the management approach (predictive, hybrid, or agile) to the specific environment in which the project unfolds, acknowledging that there is no universally valid solution.

Among the most relevant contributions in this field are the proposals by Shenhar and Dvir [46], who developed the Diamond Model based on four key dimensions: novelty, technology, complexity, and pace. Although not an operational tool, it explicitly introduces complexity as a structuring axis for approach selection. Boehm and Turner [26], for their part, propose a model based on balancing agility and discipline, using factors such as team size, product criticality, organizational experience, and requirements stability. In their framework, complexity and risk are present, albeit distributed across several criteria.

The Project Management Institute reinforces this logic through the Agile Practice Guide, which offers a set of recommendations for choosing among agile, hybrid, or predictive approaches based on variables such as environmental volatility, clarity of requirements, and degree of customer involvement. Similarly, authors such as Cobb [47] and Conforto [48] emphasize the need to analyze the nature of the project and its environment beforehand to determine the most appropriate approach. In all these cases, complexity is treated as a variable to be considered, although it is rarely operationalized rigorously or combined explicitly with risk. In all these cases, complexity is treated as a relevant variable, but it is rarely operationalized rigorously or combined explicitly with risk

Other works, such as those by Pellegrinelli [49] and Mikoba [50], point in the same direction: context and environmental variability should shape methodological choices. However, none of these models proposes a structured tool that jointly integrates both the structural dimension (initial conditions) and the dynamic dimension (changes during execution).

Despite their differences, these methods share several common traits. First, all recognize that approach selection must be contextual rather than prescriptive. Second, most introduce complexity as an element of analysis, though with varying levels of formalization. Finally, they agree that projects facing high uncertainty or rapidly changing environments tend to benefit more from adaptive or hybrid approaches.

The main differences lie in three key aspects: the level of operability of the method (ranging from conceptual models to applicable tools), the degree of formalization of the complexity variable, and whether risk is explicitly included as an element of analysis. In this respect, the model of Boehm and Turner [26] and the framework of Shenhar and Dvir [46] are the closest to integrating both aspects, albeit still only partially.

Although complex projects are developed in very different domains and complexity manifests itself differently depending on the context, they all share certain common structural and dynamic patterns, such as interdependence, non-linearity, or the emergence of emergent behaviors. These patterns can be analyzed consistently through network-based metrics, providing a transversal analytical basis for complexity and risk assessment across different sectors.

In this context, a clear gap remains between the conceptual recognition of complexity in projects and the availability of operational tools capable of integrating complexity and risk in a unified and systematic manner. This gap motivates the development of the methodology presented in this work, which extends the PMI Approach Suitability Tool by incorporating quantitative assessments of structural and dynamic complexity, together with systemic risk, to support selection of project management approaches in highly complex environments. This tool proposes a graphical, participatory method for project management teams, based on the evaluation of ten parameters grouped into three categories (culture, team, and project): (1) project (change, criticality, and delivery); (2) team (size, experience, and stakeholders); and (3) culture (acceptance, trust, and decision-making). Each parameter is scored on a scale from 1 to 10, where low values indicate greater affinity with agile approaches and high values with predictive approaches. Hybrid approaches, which combine flexibility and structure, lie between these extremes.

3.2. Proposed Methodology

The proposed method is based on an extension of the Project Management Institute’s Approach Suitability Tool, included in Annex 3 of the Agile Practice Guide [27]. The original tool is designed to select the most appropriate management approach (predictive, agile, or hybrid) according to ten parameters distributed across three dimensions: project, team, and culture. The modification proposed here replaces the “change” and “criticality” parameters with two new variables: total complexity (TC) and total risk (TR), both derived from complex network analysis applied to project stakeholders and risks.

This extension turns the instrument into a hybrid tool that combines qualitative assessments (team and culture dimensions) with quantitative metrics (project dimension), providing a more objective evaluation aligned with the dynamic and uncertain nature of complex projects. This integration also helps to strengthen project resilience by balancing structure and flexibility within a framework that explicitly acknowledges the non-linear nature of interactions among actors, risks, and processes.

The following subsections describe the underlying principles and the calculation model that incorporates the new complexity and risk parameters, as well as the logic of their integration into the modified PMI tool. The inclusion of total complexity (TC) and total risk (TR) responds to the need to more accurately reflect the systemic nature of complex projects.

The original “change” parameter is replaced by “total complexity,” with the aim of capturing not only the frequency or intensity of project changes, but also the structural and dynamic conditions that make such changes possible or inevitable. In line with Fellman [51] and Snowden and Boone [52], change is not considered the cause of complex behavior but its emergent consequence, deriving from systems characterized by high interdependence and non-linearity. Complexity therefore acts as an anticipatory and explanatory parameter for adaptive project behaviors from the early stages.

The replacement of the “criticality” parameter with “total risk” responds to the need for a quantifiable variable consistent with the dynamic nature of complex projects. While criticality is an abstract and difficult-to-objectify concept, total risk makes it possible to incorporate not only the probability and impact of events, but also their relational structure and temporal evolution. In line with Aven [53] and Hillson [54], risk is understood here as a systemic phenomenon capable of being amplified through interdependencies between factors and of generating emergent dynamics. This approach provides a more robust measure aligned with the model’s general logic, whereby a higher level of systemic risk translates into a lower TR score, signaling the suitability of agile or hybrid approaches.

3.2.1. Structure and Foundations of the Model

The evaluation structure defined for TC and TR preserves the internal logic of the PMI model, in which low values represent environments with high uncertainty and high values represent more predictable and stable contexts. To ensure this consistency, complexity and risk metrics are normalized using an inverse function on a 10-point scale, as defined in Expressions (3) and (4), where SC is structural complexity, DC is dynamic complexity, SR is structural risk, and DR is dynamic risk. The design penalizes high values of complexity or risk, so that a more uncertain environment produces lower TC or TR scores, in line with the logic of the original tool.

$$\mathrm{T}\mathrm{C}=10-(\mathrm{S}\mathrm{C}+\mathrm{D}\mathrm{C}),$$

(3)

$$\mathrm{T}\mathrm{R}=10-(\mathrm{S}\mathrm{R}+\mathrm{D}\mathrm{R}),$$

(4)

3.2.2. Calculation of the Parameters

The calculation of quantitative parameters forms the core of the proposed model, as it enables complexity and risk—often treated qualitatively—to be translated into numerical values that can be compared within the evaluation tool. Four variables are defined: structural complexity (SC), dynamic complexity (DC), structural risk (SR), and dynamic risk (DR). These are obtained through complex network analysis applied to the project and its environment, which makes it possible to quantify (1) the degree of interdependence between elements and (2) the system’s capacity to evolve and generate emergent behaviors. The values derived from these four variables are subsequently combined to obtain the aggregate parameters total complexity (TC) and total risk (TR), as defined by the equations above.

Structural complexity (SC) and structural risk (SR) are obtained from extended cyclomatic complexity (ECyM), applied, respectively, to the stakeholder network and the risk network. This metric, derived from McCabe’s work [44] and extended by Lassen and van der Aalst [43], quantifies system interdependence through the number of independent paths or cycles in its topology.

In organizational networks, higher ECyM values reflect more entangled and redundant structures, in which information or influence flows can circulate through multiple routes, increasing coordination and control difficulty [42].

The thresholds adopted are based on empirical criteria validated in process modeling and social network studies, where these intervals clearly delineate the transition from simple structures to densely connected configurations [55]. In medium-sized networks such as those analyzed in this study, these thresholds allow discrimination between linear or hierarchical systems, moderately interdependent networks, and tightly coupled architectures, thus providing an interpretive scale consistent with complex systems theory and the objectives of the proposed model.

The values are interpreted according to the thresholds shown in

Table 1. Interpretation thresholds for extended cyclomatic complexity (ECyM).

ECyM Range	Classification	Assigned Value	Interpretation
ECyM < 10	Low	0	Simple Network, low interdependence
10 ≤ ECyM ≤ 20	Medium	3	Moderately complex network
ECyM > 20	High	5	Highly entangled network

.

Dynamic complexity (DC) and dynamic risk (DR) are estimated indirectly from the number of nodes (N), used as an indicator of system coevolution, understood as the capacity to generate emergent behaviors and non-linear interactions among components [45]. In small-world networks, small structural variations can produce drastic changes in global dynamics [41], which explains why smaller networks exhibit greater sensitivity and adaptability. As N increases, the system tends to stabilize, reducing its capacity for emergent responses [21,56,57]. Accordingly, the thresholds N < 25, 25 ≤ N < 100, and N ≥ 100 for CD are interpreted as empirical transition zones between regimes of emergent, intermediate, and structurally stabilized behavior.

In the case of dynamic risk (DR), the upper limit is extended to N ≈ 200 due to the broader nature of risk networks, which include indirect interdependencies and external factors. This methodological decision is supported by studies on failure propagation in complex networks [58,59], which show connectivity thresholds beyond which disturbances can spread systemically. Additionally, the literature on social cohesion [60] suggests a relational saturation point around 150–200 agents, which offers a structural parallel to risk dynamics: above this size, the system gains structural stability but also a greater potential for global propagation of events, thereby justifying the different thresholds adopted for DC (

Table 2. Assessment criteria for Dynamic Complexity (DC).

Nº of Nodes	DC Value	Coevolution Level
N < 25	4	High
25 ≤ N < 100	2	Medium
N ≥ 100	0	Low

) and DR (

Table 3. Assessment criteria for Dynamic Risk (DR).

Nº of Nodes	DR Value	Coevolution Level
N < 25	4	High
25 ≤ N < 200	2	Medium
N ≥ 200	0	Low

).

The discretization and normalization of ECyM and node count N are not intended to define universal thresholds of complexity, but to operationalize transitions between different complexity regimes in a manner consistent with the decision-oriented nature of the PMI tool. The selected intervals reflect empirical patterns reported in process modeling and network studies, where ECyM values below 10 typically correspond to linear or weakly coupled structures, intermediate values (10–20) indicate moderate interdependence, and higher values (>20) characterize densely connected configurations with significant coordination challenges. Similarly, the discretization of N is grounded in the literature on coevolution and small-world networks, where changes in network size are associated with shifts between highly adaptive, intermediate, and structurally stabilized dynamic regimes. The use of a simple linear inverse normalization on a 0–10 scale ensures internal consistency with the original PMI framework, prioritizing interpretability and robustness over fine-grained metric precision.

3.2.3. Analytical Framework and Application

This study adopts the structural analysis framework proposed by Adami and Verschoore [61], who developed a three-level multi-network model for project analysis, composed of the supply network, the contractual network, and the information network. For the purposes of this research, the analysis focuses specifically on the information network, considered the most representative reflection of how the project system operates. This network makes it possible to understand the structure of relationships among actors and their influence on project governance.

The choice of this network is justified because, according to the authors, the information network contributes decisively to the effective governance of relationships, and its centralization structure is directly associated with communication control. Such control is a critical element for understanding emergent phenomena generated by interactions among different actors. This methodological approach enables the analysis of how structural properties of the information network—such as connectivity, density, and centrality—can give rise to emergent behaviors that are not predictable through isolated analysis of individual project components. Consequently, the multi-network approach provides a systemic perspective that captures the inherent complexity of projects and improves understanding of their organizational dynamics.

To ensure methodological transparency and reproducibility, a standardized procedure was applied for the construction and analysis of stakeholder and risk networks. In both cases, nodes were defined as discrete and identifiable entities: stakeholders correspond to organizational actors, institutions, or groups with decision-making capacity or influence over the project, while risk nodes represent distinct risk factors explicitly documented in the analyzed sources. Edges represent the existence of a direct relationship, interaction, or dependency between nodes, based on explicit evidence of coordination, influence, causality, or risk propagation. Data for network construction were collected through systematic document analysis, including official United Nations reports, mission mandates, evaluation studies, and peer-reviewed literature. To enhance reliability, node and edge definitions were cross-validated across multiple independent sources, and ambiguous relationships were excluded to avoid overfitting the network structure. Network metrics were computed using UCINET, and results were checked for consistency by verifying structural properties (number of nodes, edges, and strongly connected components) and by comparing outcomes with alternative network analysis tools when necessary. This standardized process allows the proposed methodology to be replicated across different project contexts while maintaining analytical coherence.

The procedure unfolds in two main stages:

• Stakeholder network. Mapping stakeholders and their interactions makes it possible to calculate extended cyclomatic complexity and key parameters such as centrality and strongly connected components (SCC), thus offering an in-depth view of project structure.
• Risk network. Risks are organized into a network structure, similar to a fault tree, in which each risk is connected according to its causality and influence. Analysis of this network helps to understand how risks interact and propagate.

Regarding implementation of the structural analysis, the method can be replicated using different software tools specialized in network analysis, such as UCINET [62], as described in [42], Gephi (Gephi Consortium, Paris, France) [63], Pajek (University of Ljubljana, Ljubljana, Slovenia) [64], or NetworkX (NumFOCUS, Austin, TX, USA) [65]. These platforms make it possible to compute metrics such as centrality, clustering coefficient, strongly connected components (SCC), and network density, as well as to visualize topology and interdependencies between nodes. In this research, UCINET was used as the reference tool for experimental verification of the model due to its robustness and widespread use in social and organizational network studies. However, the procedure described can be implemented in any of the mentioned environments or in other tools with equivalent functionality, thus ensuring the reproducibility and transferability of the proposed method.

3.2.4. Integration of the Parameters into the PMI Tool

The new tool incorporates TC and TR values in place of the original “change” and “criticality” parameters, while keeping the other dimensions (team and culture) unchanged. The project dimension is therefore composed of the parameters listed in

Table 4. Parameters of the “Project” dimension.

Dimension	Parameters	Nature	Source
Project	Total Complexity (TC)	Quantitative	Stakeholder network analysis
Project	Total risk (TR)	Quantitative	Risk network analysis
Project	Delivery	Quantitative	Original PMI assessment

. The combination of qualitative and quantitative variables gives the tool greater explanatory and predictive capacity, enabling a more precise selection of the management approach most suited to the project’s conditions.

In this way, the tool not only facilitates the choice of management approach but also becomes an instrument that strengthens organizational resilience by allowing structural adaptations to be anticipated before environmental changes translate into critical project deviations.

Figure 3 shows the graphical representation of the results obtained through the proposed calculation model. It visualizes the interaction between the project, team, and culture dimensions, as well as the relative position of the values obtained for each. The diagram makes it possible to identify both the structural and dynamic complexity levels of the project, along with the associated risk level, thereby facilitating the selection of the most appropriate management approach—agile, predictive, or hybrid—according to the resulting profile: (1) a central zone (low values, between 1 and 4) associated with higher-uncertainty contexts where agile approaches are recommended; (2) intermediate zones (values between 4 and 8) where structure and flexibility are balanced, favoring hybrid approaches; and (3) outer zones (high values, between 8 and 10) corresponding to stable and predictable contexts where predictive approaches are recommended.

Figure 4 presents the complete workflow of the proposed methodology. It is structured into four main phases: (PHASE 1) stakeholder network analysis, (PHASE 2) risk network analysis, (PHASE 3) assessment of the suitability of the management approach, and (PHASE 4) evaluation of project success. Each phase integrates qualitative and quantitative parameters to compute and represent a systemic view of complexity and risk. The methodological sequence begins with the identification of actors and relationships, continues with the estimation of structural and dynamic complexity, and concludes with the selection of the most appropriate management approach (agile, predictive, or hybrid), together with result validation through performance indicators.

4. Validation

To verify the applicability and internal coherence of the proposed model, the methodology was validated through three case studies focused on United Nations peacekeeping missions: UNMISS (peacekeeping mission in South Sudan), MONUSCO (peacekeeping mission in the Democratic Republic of the Congo), and MINUSTAH (peacekeeping mission in Haiti). These operations were selected due to their high organizational complexity, contextual diversity, and the availability of documented information on their management and outcomes. Taken together, they represent highly interdependent socio-technical systems, subject to continuous change and involving multiple actors with different levels of influence and distinct objectives.

Each case was analyzed by applying the proposed method to:

• Build the stakeholder and risk networks associated with each mission. This involves analyzing the management approach, identifying and mapping conflict stakeholders from the sources consulted in each case study, constructing the network of relationships among actors, and calculating betweenness centrality using UCINET software to determine structural complexity. The same procedure is then applied to the risk network to obtain structural and dynamic risk metrics.
• Calculate the total complexity (TC) and total risk (TR) parameters.
• Determine the most suitable management approach (predictive, agile, or hybrid).
• Analyze mission success by assessing the performance of each peace operation according to eight dimensions derived from United Nations and NUPI studies [66,67,68,69]: political primacy and electoral organization, security and transition, national and international ownership, regional and international support, coherence and commitment, legitimacy and impartiality, women and children, and a people-centered approach.
• Compare the identified management approach with the mission’s actual performance, evaluated along those eight dimensions.

The empirical validation makes it possible to test the model’s logical consistency and its diagnostic capacity in highly uncertain contexts, while also providing comparative evidence of its performance across diverse environments.

4.1. Contextualization of the Case Studies

UN peacekeeping missions, as objects of study, represent a clear example of complex systems in which uncertainty, multiple decision scales, and emergent dynamics coexist. From Prigogine’s perspective [57], these missions operate far from equilibrium, in environments where small perturbations can generate profound structural changes. Kauffman [70] would place them at the edge of chaos, where self-organization and innovation are possible but unstable, and for Bar-Yam their management requires an organizational structure whose complexity is comparable to that of the environment, with distributed adaptability and multilevel coordination.

UN peacekeeping operations have evolved since 1945, adapting to changes in conflicts. During the Cold War, “first-generation” missions focused on interstate conflicts under Chapter VI of the UN Charter, based on the consent of the parties, impartiality, and minimal use of force [71]. After 1989, “second-generation” missions began to address internal conflicts and promoted the protection of human rights [72]. In 1992, Boutros-Ghali’s “Agenda for Peace” introduced “third-generation” missions, which allowed interventions without the full consent of the parties and the use of force to protect civilians and ensure humanitarian assistance, under Chapter VII. However, failures such as those in Rwanda and Bosnia exposed significant limitations [73].

The 2000 “Brahimi Report” ushered in “fourth-generation” missions, emphasizing the need for clear mandates, adequate resources, and robust rules of engagement, and integrating civilian and military components to build sustainable peace. Subsequent reform processes, such as “New Horizon” and the HIPPO Report, recommended adaptive, people-centered approaches. The 2017 “Action for Peacekeeping” (A4P) initiative sought to strengthen operations in key areas, although its implementation remains ongoing [74].

Currently, peace operations are guided by principles such as consent of the parties, impartiality, and the use of force only in self-defense. An integrated approach is promoted, combining civilian and military efforts to address the complexities of modern conflicts. This requires clear mandates, effective coordination, and continuous adaptation to changing contexts to achieve lasting peace [75].

Measuring the success of peace missions is complex and contested, since contemporary conflicts are complex systems in coevolution with multiple interacting actors. Complexity theory offers a framework for better understanding these dynamics, highlighting the importance of adaptive approaches and a focus on societal resilience. Without a deep understanding of the system, resources may be expended without meeting the needs of the civilian population.

Assessing success in peace missions is a complex and debated issue, with no broad consensus on which parameters should be included in evaluation [36]. The inherent complexity of today’s conflicts calls for an approach that views each stakeholder as an agent within a coevolving complex system [76]. These agents may be local armed groups, governments, or elements of organized crime, all interacting and shaping conflict dynamics. The characteristics of complex systems—such as sensitivity to initial conditions, non-linearity, and self-organization—are directly applicable to the conflicts faced by peacekeeping missions [77].

Several authors argue that complexity theory has great potential for studying international conflicts [78,79]. Classical systems theory often fails to adequately address the multiple intersections and contradictions that arise among actors in a complex system [80]. The newer perspective enables a better understanding of intervention and post-conflict reconstruction policies, showing that without a deep grasp of the system, resources can be wasted without addressing the needs of the civilian population [81].

Building peace in complex conflicts therefore requires an adaptive approach centered on societal resilience [68]. The 2015 HIPPO report [82] stresses that peacebuilding is a political process that requires integrated and adaptive missions, rejecting standardized approaches in favor of context-specific adaptation [83]. Morton [23] also suggest adaptive actions for peacekeeping, including context-specific approaches, participatory methods, and iterative processes.

Measuring the success of a peace mission involves considering both strategic and tactical factors [84]. The three basic principles are consent of the parties, impartiality, and the use of force only in self-defense [85]. However, vague mandates, inadequate resources, and a lack of clarity in the application of these principles have led to the failure of many missions [86,87].

To assess the success of peace missions, this study develops an analytical framework based on eight fundamental dimensions that synthesize the principles and best practices established by both the United Nations and NUPI [67,68,69]. Although other evaluation frameworks exist [88,89], these eight dimensions capture the essential elements needed to determine whether a peacekeeping operation is achieving its objectives:

• Political primacy and electoral organization: Establishing stable political structures and peaceful electoral processes.
• Security and transition: Protection of civilians and stabilization during crises.
• National and international ownership: Ensuring that the mission is perceived as belonging to the local population.
• Regional and international support: Backing from neighboring countries and international organizations.
• Coherence and commitment: A clear mandate and active societal participation.
• Legitimacy, impartiality, and credibility: The mission must be legitimate, impartial, and credible to all parties.
• Women and children: Protection and empowerment of the most vulnerable groups.
• People-centered approach: Prioritizing individual well-being and security.

The three missions selected for the case studies (UNMISS, MONUSCO, and MINUSTAH) are fourth-generation operations that intervene in complex internal conflicts, often without prior peace agreements, and combine civilian and military components. Civilian personnel require expertise in advanced project management systems to adapt to highly unstable and rapidly changing environments. Moreover, these missions have been active for extended periods in their respective territories, enabling a thorough assessment of their performance and a deeper understanding of the factors that affect their success and sustainability.

Developed in distinct geopolitical and social contexts, these missions bring diversity to the study, allowing an evaluation of how different management approaches influence success and enabling the testing of the proposed tools and methodologies across various scenarios. The analysis focuses on measuring the success of these missions, identifying the most suitable approach for managing a peace operation, and exploring the relationship between that approach and the outcome achieved, considering the specific variables of each context.

The challenges in managing UN peacekeeping missions stem primarily from the complexity and high instability of the environments in which they are deployed. The diversity of actors involved—such as civilian and military forces with specific roles that must be effectively coordinated—and the constantly changing conditions of security, local interests, and political dynamics make it difficult to implement a stable management approach. These complications may arise due to external factors, such as opposition from certain groups, or internal factors, such as communication limitations and coordination issues among mission components.

4.2. Case Study 1: UNMISS, South Sudan

The United Nations Mission in South Sudan (UNMISS) was established immediately after the conclusion of the United Nations Mission in Sudan (UNMIS), which had overseen South Sudan’s transition to independence from Sudan. Using the integrated CPAS approach described in the Brahimi Report, UNMISS aimed to support the establishment of the new state following decades of conflict between the predominantly Arab north and the animist and Christian south. Key agreements—such as the 1972 Addis Ababa Peace Agreement, the 2002 Machakos Protocol under IGAD auspices, and the 2005 Comprehensive Peace Agreement, which established a period of autonomy and equitable distribution of oil revenues—were fundamental in this process.

Despite initial progress, a civil war erupted in South Sudan in 2013 between government forces and rebel factions, forcing UNMISS to shift its focus toward the protection of civilians. Although an agreement to resolve the political dispute was signed in 2015, fighting resumed in 2016. A new agreement was reached in 2018 under IGAD’s mediation, remaining in effect despite continued skirmishes in the region. Throughout these events, the UN adapted its missions—from UNAMIS to UNMIS and finally UNMISS—to address the evolving challenges in Sudan and South Sudan.

To study the context and evolution of the situation in Sudan and South Sudan, the following sources were consulted in addition to UNMISS reports [90,91,92,93,94,95,96,97].

4.2.1. Phase 1: Stakeholder Network Analysis

Based on the consulted sources, the main stakeholders in the conflict were identified, and a relational network was constructed to analyze betweenness centrality using UCINET to determine structural complexity. Figure 5 shows the stakeholder network identified in the UNMISS mission and the betweenness centrality of each actor. This representation makes it possible to visualize stakeholder relationships and identify the nodes with the greatest influence on project dynamics.

Figure 6 presents the data on nodes, edges, and strongly connected components (SCC) derived from the network analysis.

Using Expression (1) defined in the methodology section, the resulting ECyM value for the stakeholder network is 13, corresponding to a medium level of structural complexity. Applying Expression (3), the total complexity (TC) equals 3, classifying the mission as a complex project.

4.2.2. Phase 2: Risk Network Analysis

Figure 7 presents the risk network identified for the UNMISS mission, along with betweenness centrality for each node. This representation illustrates the structure of interdependencies among risk factors and highlights those elements most influential in the propagation of critical project events.

Using the risk network and applying Expression (1), the resulting ECyM value is 1, indicating a simple risk structure with low interdependence. According to the criteria defined in Expression (4), total risk (TR) is calculated as 6, classified as moderate risk. The results are summarized in Figure 8 and Figure 9, which show the suitability assessment and the modified management-approach diagram for the UNMISS mission.

4.2.3. Phase 3: Evaluation of the Project Approach

Findings suggest that a hybrid approach is the most suitable for this mission for three main reasons: the size and experience of the team; the level of trust and acceptance of a hybrid approach within the team, and the presence of moderate risk.

Within the project dimension, the three perspectives call for different approaches: complexity is best addressed through an agile approach due to the radical changes triggered by the outbreak of civil war; risk management is best handled through a hybrid approach given the need to establish protection sites and conduct control tasks typical of traditional missions; and delivery, given the impossibility of completing the mission, requires a more conservative predictive approach.

4.2.4. Phase 4: Mission Success Analysis

The UNMISS mission in South Sudan faced severe political, social, and economic challenges, including an interethnic civil war and widespread destruction of infrastructure. Its primary objective—to establish governance structures and organize elections—failed due to the ongoing conflict. However, the mission achieved success in protection and stabilization by creating safe zones for more than 200,000 displaced civilians and by implementing a decentralized structure that encouraged local participation.

Despite these advances, regional and international support declined, and there were failures in coordination with the central government, affecting coherence and participation. Efforts to reduce violence and address sexual violence against women were insufficient. Moreover, the mission did not effectively implement a people-centered approach, often prioritizing strategy over direct engagement with the population.

Nevertheless, this decentralized structure and the involvement of local actors contributed to strengthening institutional and community resilience in the face of violence and displacement.

Figure 10 presents the evaluation of the eight success dimensions, indicating the degree of achievement for each one and the observations that support the assessment.

4.3. Case Study 2: MONUSCO, Democratic Republic of the Congo

The massive flight of more than one million Tutsis following the massacres perpetrated by Hutus in Rwanda in 1993 destabilized the already fragile political and social situation in the former Zaire. The alliance between the political opposition to Mobutu’s regime and the displaced Tutsis, supported by Uganda, triggered a civil war that culminated in 1997 with the fall of Mobutu. The rebels seized Kinshasa and renamed the country the Democratic Republic of the Congo (DRC). However, another civil war broke out in 1998: forces loyal to President Laurent-Désiré Kabila—supported by Angola, Chad, Namibia, and Zimbabwe—faced opponents backed by Rwanda, Burundi, and Uganda, in what came to be known as Africa’s First World War.

In response to the resulting humanitarian crisis, the UN adopted Resolution 1234/99, calling for an end to violence and the establishment of peace. That same year, the Lusaka Agreement was signed and the MONUC mission was created, focusing on civilian protection and humanitarian assistance. Despite multiple agreements, such as the Sun City accords in 2002 and the Goma agreement in 2009, and the transition from MONUC to MONUSCO in 2010 to help stabilize peace and support elections, the country has continued to experience recurrent conflicts, especially in the east. In 2013, the UN reinforced the mission in response to renewed uprisings. After international pressure, President Joseph Kabila stepped down in 2018, but the DRC remains in a fragile peace, particularly in the eastern regions bordering Uganda.

To study the context and evolution of the situation in the Democratic Republic of Congo, the following sources were consulted in addition to MONUSCO reports [98,99,100,101,102,103,104].

4.3.1. Phase 1: Stakeholder Network Analysis

Based on the consulted sources, the main stakeholders in the conflict were identified, and the resulting network and betweenness centrality analysis were conducted using UCINET to determine structural complexity. Figure 11 illustrates the stakeholder network and betweenness centrality for the MONUSCO mission.

Figure 12 presents the data on nodes, edges, and strongly connected components (SCC) derived from the network analysis.

Using Expression (1) defined in the methodology section, the stakeholder network obtains an ECyM value of 20, corresponding to a medium-high level of structural complexity. According to the established classification criteria, and applying Expression (3), total complexity (TC) equals 5, defining the mission as a moderately complex project.

4.3.2. Phase 2: Risk Network Analysis

Figure 13 shows the risk network built from the analyzed sources, as well as the betweenness centrality of each node. This representation helps identify the most influential risks and causal relationships among them.

Using the risk network and Expression (1), the resulting ECyM value is 1, indicating a simple risk structure with low interdependence. Applying the coevolution criteria and Expression (4), total risk (TR) is 5, classified as moderate risk.

4.3.3. Phase 3: Evaluation of the Project Approach

The results of the analysis are summarized in Figure 14 and Figure 15, which, respectively, show the suitability assessment and the modified approach-selection diagram for the MONUSCO mission. These representations indicate that the most suitable approach corresponds to a hybrid configuration.

4.3.4. Phase 4: Mission Success Analysis

The MONUSCO mission adopted a fully hybrid approach, expanding its size to confront rebel groups in the eastern part of the country. Complexity falls within the hybrid range due to the shift from a political mission to a more managerial one. Risks—including the Ebola epidemic and armed clashes—were controlled thanks to military support that facilitated immediate peace talks.

When evaluated across the eight proposed dimensions, MONUSCO obtained mixed results. In terms of political primacy and electoral organization, it achieved early successes through peace agreements and electoral processes, but its political influence later declined due to insufficient cooperation from the Congolese government. In protection and stabilization, the mission implemented disarmament and repatriation processes, but instability persisted in the east because of armed groups and resource limitations.

Local and national ownership was a failure, as the government ignored reforms and the population perceived the mission as an external intervention. Regional and international support improved conditions compared to previous years, although interference from neighboring countries complicated the process. Coherence and commitment weakened due to frequent leadership changes and the transition from a political to a managerial focus, undermining relations with civil society and the government.

Regarding legitimacy, impartiality, and credibility, the mission achieved early successes, particularly in public health, but lost political alignment over time. Concerning women, peace, and security, it made progress by promoting female empowerment and addressing sexual violence. However, the people-centered approach was limited by insufficient resources and the inability to prevent the proliferation of armed groups, leading the population to view the mission as disconnected from their needs.

Overall, MONUSCO contributed to reducing violence and improving certain social conditions but faced significant obstacles due to lack of governmental cooperation and strategic challenges. Despite persistent difficulties, the mission demonstrated notable operational resilience, maintaining its response capacity during simultaneous crises—such as the Ebola epidemic—without collapsing its functional structure.

Figure 16 presents the evaluation of the eight success dimensions, indicating the degree of achievement for each one and the observations that support the assessment.

4.4. Case Study 3: MINUSTAH, Haiti

The United Nations has been involved in Haiti since the 1990s through several missions—including UNMIH, UNSMIH, UNTMIH, and MINOPUH—aimed at stabilizing the country following the fall of the Duvalier dictatorship in 1986. Despite attempts to establish democratic governments, Haiti experienced persistent political instability marked by coups, institutional crises, and violence carried out by paramilitary groups known as “tonton macoutes”.

In response to the armed conflict and institutional crisis that followed President Aristide’s exile in 2004, the UN Security Council established the United Nations Stabilization Mission in Haiti (MINUSTAH) through Resolution 1542. The mission sought to stabilize the country, support a political transition in a secure environment, and promote a liberal democratic framework. After the devastating 2010 earthquake—which caused 220,000 deaths and triggered a cholera outbreak—MINUSTAH intensified its efforts to address the humanitarian and economic crisis. In October 2017, the UN concluded MINUSTAH, determining that Haiti was prepared to advance with greater autonomy, and established a smaller follow-on mission, MINUJUSTH, to continue strengthening the country’s judicial system.

To study the context and evolution of the situation in Haiti, the following sources were consulted in addition to MINUSTAH reports [105,106,107,108,109,110,111].

4.4.1. Phase 1: Stakeholder Network Analysis

Based on the consulted sources, the main stakeholders in the conflict were identified, along with the network they form and its betweenness centrality, analyzed using UCINET to determine structural complexity. Figure 17 shows the stakeholder network and the betweenness centrality levels identified for the MINUSTAH mission, allowing visualization of the power structure and influence among key actors.

Figure 18 presents the data on nodes, edges, and strongly connected components (SCC) derived from the network analysis.

Applying Expression (1) defined in the methodology section, the stakeholder network yields an ECyM value of 13, corresponding to a medium level of structural complexity. According to the established criteria and applying Expression (3), total complexity (TC) equals 3, classifying the mission as a complex project.

4.4.2. Phase 2: Risk Network Analysis

Figure 19 presents the risk network constructed from the analyzed sources, showing betweenness centrality for each node and the interdependencies among risk factors. This representation helps identify the most influential risks in the mission’s systemic dynamics.

Applying Expression (1), the risk network yields an ECyM value of 2, indicating a simple risk structure with low interdependence. According to the coevolution criteria and Expression (4), total risk (TR) reaches a value of 6, corresponding to moderate risk.

4.4.3. Phase 3: Evaluation of the Project Approach

The results of the analysis are summarized in Figure 20 and Figure 21, which present, respectively, the suitability assessment and the modified approach-selection diagram for the MINUSTAH mission.

The results indicate that the most appropriate configuration corresponds to a hybrid approach with some proximity to a predictive approach, largely due to the need for police forces to take firm action against armed gangs and the severe hardship faced by the civilian population. In terms of complexity, the mission falls at the agile end of the spectrum, largely due to the rapidly shifting conditions that shaped mission operations.

4.4.4. Phase 4: Mission Success Analysis

The assessment of MINUSTAH in Haiti across the eight dimensions shows mixed results. Politically, the mission facilitated elections in 2006, 2010–2011, and 2016, helping restore constitutional order after periods of coups and interim governments. In protection and stabilization, despite efforts to rebuild institutions and provide humanitarian aid—especially after the devastating 2010 earthquake—the mission struggled to reduce violence, corruption, and the influence of armed gangs, which continued to destabilize the country.

Regarding national and local ownership, although the population appreciated the humanitarian assistance, the mission was criticized for failing to achieve significant progress in security and economic development and for what many perceived as excessive foreign intervention—particularly from the United States—seen as driven by external interests. Coherence and participation were affected when participating countries used the mission to intervene directly, potentially deviating from the UN mandate.

In terms of women, peace, and security, reporting of sexual violence increased, and initiatives were introduced to protect women’s rights, but significant challenges persisted in gender equality and protection against violence. The people-centered approach was limited by natural disasters, political instability, and extreme poverty, preventing sustained progress.

Since MINUSTAH’s closure in 2017, conditions in Haiti have deteriorated, with rising poverty, insecurity, and political crisis—highlighted by the assassination of President Jovenel Moïse in 2021—leading some analysts to consider the country a failed state. The reconstruction efforts after the 2010 earthquake illustrated both the importance and the limitations of societal resilience in contexts where institutional fragility and external dependency constrain the long-term sustainability of the project.

Figure 22 presents the evaluation of the eight success dimensions, indicating the degree of achievement for each one and the observations that support the assessment.

5. Discussion

The primary advantage of the proposed methodology lies in its ability to deliver a quantitative and systematic assessment of complexity and risk in complex projects. From this perspective, resilience can be interpreted as an emergent indicator of the balance achieved between complexity, risk, and adaptability. The more balanced this systemic triangle is, the greater the project’s resilience and its ability to sustain outcomes under changing conditions.

Within this conceptual framework, the relationship between project complexity, systemic risk, and the selection of the management approach can be interpreted coherently through established theories of project and organizational resilience. In these theoretical perspectives, resilience is not limited to the capacity to absorb disturbances and recover performance but is understood as the ability of socio-technical systems to anticipate disruptions, adapt governance structures, and reorganize under changing conditions. Complexity thus acts as a structural and dynamic driver of emergent risk, while the choice of management approach constitutes a key mechanism through which adaptive capacity and, ultimately, project resilience are shaped.

By integrating into the complexity analysis (1) stakeholder networks, (2) risk networks, (3) the cyclomatic complexity metric, and (4) coevolution parameters, the methodology provides a more detailed understanding—expressed through precise and objective results—of both the structure and dynamics of the project. By combining qualitative and quantitative approaches, it establishes a robust foundation for selecting the most appropriate management approach (predictive, agile, or hybrid) according to the specific characteristics of the project. For this reason, the methodology overcomes the limitations of tools that rely exclusively on qualitative assessments. It thus serves as a useful instrument for improving planning, adaptation, and effective project management in contexts of high uncertainty.

The adaptation of the Agile Suitability Filter from Appendix X3 [27] introduces two conceptually relevant modifications: replacing the parameter “change” with complexity and “criticality” with risk. These adjustments enable a more detailed and meaningful evaluation for projects operating in uncertain environments. They also broaden the analytical scope of the original tool by incorporating factors related to the dynamic and structural nature of complex projects. While the original model focused on the frequency of changes and the level of criticality, the inclusion of complexity as a key aspect allows for the evaluation of multiple interactions and internal dynamics within the project. Likewise, the integration of risk adds an additional layer of analysis that directly addresses the probability and impact of uncertain events. These adjustments strengthen the methodology’s capacity to identify an appropriate management approach—predictive, hybrid, or agile—by aligning project evaluation with the characteristics of its context, thereby increasing the likelihood of success by better adapting to operational realities.

Building on this perspective, the proposed methodology explicitly operationalizes the link between complexity, risk, and management approach in such a way that resilience can be addressed during the early planning stages of the project, rather than being treated solely as a property observed ex post after disruptions occur. By translating structural and dynamic indicators of complexity into risk profiles and explicit management-path decisions, resilience becomes a property that can be deliberately embedded through design choices related to governance, coordination mechanisms, and methodological configuration. In this way, resilience emerges not only from reactive responses to disruptions but also from informed decisions made prior to project execution.

The application of the proposed methodology to the three case studies (UNMISS, MONUSCO, and MINUSTAH) demonstrates its proper functioning by enabling a coherent and context-sensitive evaluation of each mission. By analyzing stakeholder and risk networks in each case, the methodology accurately quantified both complexity and risk. For example, in UNMISS, the high structural and dynamic complexity effectively reflected the mission’s volatile environment, leading to the identification of a hybrid approach as the most suitable. This conclusion aligns with the observed reality, in which the mission had to continuously adapt to changing conditions in South Sudan’s conflict. Similarly, in MONUSCO and MINUSTAH, the methodology highlighted key differences in their approaches and challenges. In MONUSCO, moderate complexity and controlled risk indicated that a hybrid approach was appropriate, consistent with the mission’s evolution toward greater management and stabilization functions. In MINUSTAH, high complexity and moderate risk suggested the need to combine agile and predictive approaches, mirroring Haiti’s fluctuating political and social dynamics. The results confirm the validity of the tool by providing recommendations that match the circumstances of each mission and underscore its practical usefulness for planning and effectively managing peace operations in complex environments.

Despite its advantages, the proposed methodology presents several limitations that should be explicitly acknowledged. First, the construction of stakeholder and risk networks depends to a large extent on data availability and quality. In the analyzed peacekeeping missions, information was drawn from official United Nations reports and independent evaluations; however, in contexts where documentation is fragmented, outdated, or constrained by political factors, certain relationships or risk interactions may be underrepresented. This limitation directly affects the completeness of the risk network and may lead to conservative estimates of structural or dynamic complexity and risk.

Second, although the methodology is applicable to projects of different sizes, cases with a limited number of nodes may introduce specific biases. In small networks, indicators related to dynamic complexity and coevolution tend to exhibit higher variability, which may amplify the apparent sensitivity of the system to local interactions. While this behavior is consistent with complex systems theory, it requires careful interpretation when comparing projects with markedly different network sizes.

Moreover, the proposed approach requires a certain level of professional expertise in both project management and network analysis, as well as additional analytical effort compared to purely qualitative tools. Possible directions for improvement include the development of semi-automated procedures for network construction, the integration of digital trace data to complement documentary sources, and the progressive calibration of thresholds using larger and cross-sector datasets. These improvements would reduce analyst dependency, increase robustness, and facilitate the adoption of the methodology in operational project management environments.

Although the empirical validation presented in this study focuses on United Nations peacekeeping missions, the analytical logic of the proposed approach is not domain-specific. The methodology can be transferred to other complex project environments—such as mega-infrastructure projects, emergency response operations, or large engineering programs—by redefining nodes as project stakeholders, organizational units, or technical subsystems, and risks as technical, organizational, regulatory, or environmental threats. Under these reinterpretations, the network-based representation and the associated complexity and risk metrics remain applicable, enabling comparable assessments of interdependence, risk propagation, and management-path suitability across diverse contexts.

A future line of research could consist of systematically reformulating certain existing parameters in the PMI approach suitability tool to more accurately capture the characteristics inherent to complex systems. Following the logic of the modifications already introduced—where “change” is reformulated as “complexity” and “criticality” as “total risk”—other parameters could be reconceptualized from the perspective of complex systems theory. For instance, the parameter “Experience” in the Team dimension could be reformulated as “Adaptive Capacity” to assess not only prior experience but also the ability to navigate emerging uncertainty. Similarly, “Decision-Making” in the Culture dimension could be redefined as “Emergent Autonomy,” capturing the system’s ability to generate self-organized responses. The parameter “Stakeholders” could become “Systemic Connectivity” to evaluate the quality and density of information networks among actors. This conceptual revision would preserve the operational structure of the tool while aligning its components with theoretical principles of emergence and self-organization in complex systems.

Other lines of work may include applying the methodology to projects of different natures (such as industrial, technological, or infrastructure projects), as well as extending it toward multi-network models that integrate not only stakeholders and risks but also resources, processes, and teams. Additionally, future research could examine the incorporation of advanced risk analysis tools—such as Petri net models, neural networks, or artificial intelligence systems—to weight probabilities, detect emerging patterns, and strengthen project resilience in VUCA environments (Volatile, Uncertain, Complex, and Ambiguous).

6. Conclusions

This study develops a methodology to determine the most appropriate management approach for the successful execution of complex projects, applying it initially to United Nations peacekeeping missions as a validation context. Based on the Project Management Institute’s Approach Suitability Tool, the proposal expands its scope through the incorporation of quantitative metrics of complexity and risk derived from network analysis. The method enables the assessment of project complexity and risk by analyzing networks and evaluating both the structure and dynamics of stakeholder and risk networks. By incorporating structural (cyclomatic) complexity and the coevolution parameter g, the methodology provides a deeper understanding of how project elements interact and evolve. Systemic risk is addressed through the combined evaluation of structural and dynamic risk within the risk network, allowing a clearer view of how risks may affect the project.

The application of the methodology to the three case studies of United Nations peacekeeping missions (UNMISS, MONUSCO, and MINUSTAH) allowed verification of its coherence and practical usefulness. In the cases examined, the approach suitability tool helped identify the most appropriate management approach, with hybrid approaches proving effective in complex and volatile environments. The analysis of mission success further served to verify whether the selected approach was appropriate, although this component may vary in other projects depending on their specific context.

The proposed methodology constitutes a robust tool for managing complex projects, enabling the adaptation of the management approach to each context’s structural, dynamic, and risk-related characteristics. By aligning the management approach with the specific attributes of each project, it offers a conceptual and practical foundation for improving planning, adaptability, and resilience across diverse sectors. Ultimately, resilience represents the attribute that transforms complexity and risk into organizational learning. Projects capable of absorbing shocks, reorganizing, and maintaining their purpose are resilient systems—and these are precisely the ones capable of generating.