Next Article in Journal
Persistence and Resilience in Smart/Hybrid Working Practice: A Gender Evaluation in Public Sector
Previous Article in Journal
How Does AI Transform Cyber Risk Management?
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Systems
Volume 13
Issue 10
10.3390/systems13100836
systems-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Research on the Selection of Multi-Agent Interaction Modes in Complex Product R&D Networks Under Disruption Events

by
Songsong Cheng
1 and
Qunpeng Fan
2,*
1
School of Accounting, Guangzhou College of Commerce, Guangzhou 511363, China
2
School of Economics, Management and Law, Jilin Normal University, Siping 136000, China
*
Author to whom correspondence should be addressed.
Systems 2025, 13(10), 836; https://doi.org/10.3390/systems13100836
Submission received: 10 August 2025 / Revised: 15 September 2025 / Accepted: 22 September 2025 / Published: 24 September 2025
(This article belongs to the Section Complex Systems and Cybernetics)
Browse Figure
Versions Notes

Abstract

The negative impact of disruption events (i.e., departure of key personnel, geopolitical conflicts) on normalized economic operations is self-evident. How complex product R&D networks maintain resilience under disruptions is an unavoidable management issue. Existing research has primarily focused on the effects of inter-organizational interactions under disruption events, while paying little attention to the diversity of multi-agent interactions in complex product R&D network and the deeper logic embedded within them. Combining the attention-based view and sequential attention logic, we take the resilience of complex product R&D networks as the dependent variable to study the selection effects of multi-agent interaction modes in complex product R&D networks under disruption events. A hierarchical linear regression was conducted on 427 questionnaire samples from the high-end equipment manufacturing industry. Empirical results show that complex product R&D networks are prone to two multi-agent interaction modes under disruption events: the focal firm-dominated mode and the multi-agent co-creation mode. Compared with the latter, focal firms show a stronger preference for the former mode in disruptions. Both multi-agent interaction modes contribute to enhancing network resilience. Further research shows that under first-order disruption events (i.e., departure of key personnel), both focal firm-dominated mode and the multi-agent co-creation mode serve as effective resilience governance strategies for complex product R&D networks, with the multi-agent co-creation mode being better than the focal firm-dominated mode. Under second-order disruption events (i.e., geopolitical conflicts), only the multi-agent co-creation mode emerges as an important mechanism for resilience governance of complex product R&D networks. The findings provide a theoretical foundation for complex product R&D networks to select the optimal multi-agent interaction mode to enhance resilience and effectively respond to disruption events.

1. Introduction

Complex products refer to high-technology, high-value, single-unit or small-batch customized, engineering-intensive large-scale products, systems, or infrastructure [1], serving as important symbols of the country’s advanced industrial capabilities and technological innovation. Due to the rich component technologies and the intricate knowledge elements involved in complex products [2], their development cannot be achieved by a single enterprise. As the leaders in complex product development, focal firms need to form dynamic R&D networks with diverse partners through formal or informal relationships such as contracts, collaborative agreements, and personal ties, that is, to build complex product R&D networks [3]. Although such networks foster innovation, they are also inherently vulnerable [4]. Disruption events such as task restructuring, the departure of key personnel, or abrupt market shifts can destabilize established collaboration modes, delay complex product R&D progress, and increase costs. For instance, the Boeing 787 Dreamliner program experienced significant delays due to coordination failures across its global R&D network. The example underscores a pressing question: how can complex product R&D networks maintain resilience when confronted with disruption events?
This issue has garnered increasing scholarly attention under the concept of resilience governance [5], defined as the strategies focal firms adopt to anticipate, recover, and adapt from disruption event within complex product R&D networks. Prior research emphasizes that the effectiveness of resilience governance largely depends on the multi-agent interaction among complex product R&D network. Two primary multi-agent interaction modes have been identified [6]: (1) the focal firm-dominated mode, in which focal firms place their own development at the center of gravity, actively structuring interactions with other participants through dominant behaviors and thereby exhibiting centralized characteristics; and (2) the multi-agent co-creation mode, in which focal firms emphasize the holistic development of the complex product R&D network, guiding participants to engage in equitable resource exchanges and other interactions under shared visions, goals, and value frameworks, thereby exhibiting decentralized characteristics. Despite these insights, two critical gaps remain. First, it is still unclear how disruption events influence the selection of two multi-agent interaction modes in complex product R&D networks. Second, the implications of these selections for complex product R&D network resilience have not been sufficiently examined. Addressing these gaps is essential not only for advancing theory but also for providing guidance to managers operating in VUCA (volatile, uncertain, complex, ambiguous) environments.
The selection of multi-agent interaction modes in complex product R&D networks is essentially a process of allocating attention across decision-making activities [7]. Consequently, to address these questions, the study adopts the Attention-Based View (ABV) and sequential attention logic as its theoretical foundation. ABV conceptualizes attention as a scarce organizational resource [8]. Sequential attention logic extends this view by emphasizing the distributive and selective nature of attention, which shapes decision-making by determining which issues are prioritized [9]. It highlights that firms cannot process all stimuli simultaneously and must focus on certain priority problems while neglecting others. Taken together, these perspectives offer a coherent framework for explaining how focal firms allocate attention and select between the focal firm-dominated mode and multi-agent co-creation mode under disruption events, thereby generating differentiated complex product R&D network resilience governance outcomes.
Moreover, Anderson and Lewis [10] divided disruption events into three types: departure of key personnel, task and technological changes, and extreme disaster events, further classifying them into two levels: first-order disruptions and second-order disruptions. First-order disruptions focus on internal organizational changes, such as member turnover, technological changes, and task restructuring. Second-order disruptions focus on external environmental turbulence, including market structural adjustments, business model innovations, and extreme events. The distinct disruption contexts created by these two disruption event levels may shape the relative effectiveness of focal firm-dominated and multi-agent co-creation modes in different ways. Thus, a nuanced understanding of how disruption event levels influence the network resilience effects of multi-agent interaction modes is critical for advancing resilience governance in complex product R&D networks.
The marginal contributions are reflected in the following three aspects: First, by introducing a multi-agent interaction perspective to examine resilience governance in complex product R&D networks under disruption events, the study extends the research framework of network resilience governance and enriches the contextual applications of multi-agent interaction research. Second, by exploring the diversity and preference of multi-agent interaction mode selection under disruption events—through the dimensions of the focal firm-dominated mode and the multi-agent co-creation mode, the study breaks away from the limitations of static discussions on firms’ attention allocation and inspires future research to focus on the dynamics of attention allocation. Third, by comparatively analyzing the governance effects of focal firm-dominated and multi-agent co-creation modes on network resilience under disruption events, the study helps address the complex issue of multi-agent interaction mode selection faced by complex product R&D networks.

2. Theoretical Foundations and Research Hypotheses

2.1. Attention-Based View (ABV)

ABV holds that firms manage their limited attention based on three interrelated principles: context, focus, and allocation [7]. The context principle refers to firms determining their focus on specific issues and solutions based on their contextual environment [11]. The focus principle refers to the fact that the issues and solutions a firm prioritizes influence its behavioral decisions, i.e., attention allocation strategies. The allocation principle refers to how attention is distributed across specific issues and alternative solution [12]. Therefore, even for the same issue under the same context, differences in firms’ attentional focus and allocation strategies lead to different behavioral results. Following the logic, the selection of multi-agent interaction modes in complex product R&D networks is a process in which focal firms structurally allocate attention based on contextual factors such as operational, network, and task environments. Thus, ABV provides a framework to explain: which multi-agent interaction modes focal firms prioritize in disruption contexts, what attention allocation strategies they adopt, and what the expected outcomes of these strategies are.

2.2. Sequential Attention Logic

Sequential attention logic is a core behavioral mechanism within ABV framework. According to sequential attention logic, attention has the characteristics of distributivity and selectivity. Firms can allocate attention to multiple schemes simultaneously but differ in the sequence and priority assigned to each [9]. In other words, firms prioritize attention to solutions perceived as more effective or competitive in a specific context, while paying less attention to less impactful alternative [13]. Therefore, under disruptions, focal firms in complex product R&D networks may allocate attention simultaneously and unevenly in different multi-agent interaction modes. While limited attentional resources are competed for by various interaction activities, determining which multi-agent interaction mode receives priority requires in-depth analysis guided by sequential attention logic.

2.3. Disruption Events and Multi-Agent Interaction Modes in Complex Product R&D Network

As the key actors in complex product R&D networks, focal firms have strong influence and control over the whole network [14]. Their interactions with other participants of complex product R&D networks form different multi-agent interaction modes: the focal firm-dominated mode and the multi-agent co-creation mode [6]. The focal firm-dominated mode refers to focal firms taking their own survival and development as the center of gravity, actively structuring interactions with other network participants through dominant behaviors, showing centralized characteristics focal firms. The multi-agent co-creation mode means that focal firms take the holistic development of the complex product R&D networks as the center of gravity, guiding other participants to engage in equitable resource exchanges and other interactions under shared visions, goals, and value frameworks, showing decentralized characteristics.
Disruption events disrupt the collaborative inertia and interaction dependency paths within complex product R&D networks [15], thereby stimulating interaction motivations among network multi-agents. According to the ABV, firms’ attentional focus and allocation strategies depend on their specific situational contexts [11]. In disruption contexts, focal firms as rational economic actors prioritize their survival and interests [16], which leads them to focus and allocate attention on self-oriented issues and solutions. Therefore, after the disruption event occurs, focal firms consciously acquire and orchestrate resources aligned with their needs during interactions with other participants in the complex product R&D networks [17], continuously strengthening their dominance in resource allocation and benefit distribution. Finally, the multi-agent interaction of complex product R&D networks presents the characteristics of focal firm centralization, that is, the focal firm-dominated mode emerges.
According to the distributed nature of attention, firms can allocate attention to multiple schemes simultaneously [13]. As central actors in complex product R&D networks, focal firms should undertake the overall network scheduling in the face of disruption events. Thus, beyond self-interest, focal firms also allocate attention to the collective interests of network participants in complex product R&D network. This means that during interactions, focal firms emphasize shared value propositions and aligned action goals while establishing equitable dialogue and trust-based mechanisms to promote resource sharing and value co-creation among multi-agents in the complex product R&D [18]. Thus, multi-agent interactions of complex product R&D networks present decentralized characteristics, that is, the emergence of multi-agent co-creation mode.
Further, according to sequential attention logic, focal firms of complex product R&D networks may unbalanced allocate attention between the focal firm-dominated mode and multi-agent co-creation mode during disruptions. The formulation of the priority attention allocation strategy of the focal firms depends on the relative utility of the two multi-agent interaction modes in disruption contexts. The focal firm-dominated mode, characterized by dominant behaviors meeting to focal firms’ developmental needs, requires relatively small attentional resources while offering significant leverage effect. Thus, the focal firm-dominated mode is one of the important strategies for focal firms to control disruption events or use disruption events in the short term. The multi-agent co-creation mode generates sustained resource interaction based on multi-agents’ collective development in complex product R&D networks. However, coordinating multi-agents to reach a co-creation consensus consumes a lot of attentional resources of focal firms [19]. It has the characteristics of long cycle, high cost and high risk, and it is difficult to promote focal firms to respond quickly to disruption events or develop opportunities derived from disruption events. Thus, when facing disruptions, focal firms in the complex product R&D network exhibit stronger inclinations toward adopting the focal firm-dominated mode, with weaker motivations to develop the multi-agent co-creation mode.
Thus, we propose the following hypotheses:
H1. 
Complex product R&D network under disruption events is prone to two multi-agent interaction modes: the focal firm-dominated mode and the multi-agent co-creation mode.
H2. 
Compared to the multi-agent co-creation mode, the focal firm-dominated mode is more likely to emerge in complex product R&D networks under disruption events.

2.4. Multi-Agent Interaction Modes and Resilience of Complex Product R&D Networks

With the aggravation of geopolitical risks and the continuous emergence of disruptive technologies, resilience governance of complex product R&D networks has become a critical concern for firms. The resilience of complex product R&D networks refers to their capacity to buffer responses, reorganize evolution, and dynamically adapt following disruption shocks, enabling self-organized recovery of network structures, functions, and interactive feedback mechanisms [20]. Although frequent multi-agent interactions emerge in complex product R&D networks under disruption events, manifesting as focal firm-dominated mode and multi-agent co-creation mode, the specific mechanisms through which these multi-agent interaction modes influence complex product R&D network resilience remain underexplored.
The key to resilience governance in complex product R&D networks during disruption events lies in multi-agent cooperation to overcome adverse disturbances. Under the multi-agent co-creation mode, focal firms facilitate interactive co-creation among network agents by shaping shared value propositions [21]. This approach not only enhances compatibility, mutual recognition, and coordinated development among multi-agents but also strengthens complex product R&D network adaptability and flexibility through resource connection during disruptions. Furthermore, it fosters strong trust mechanisms among multi-agents, reduces opportunistic behaviors, and ensures orderly complex product R&D network operations post-disruption. These studies suggest that multi-agent interaction mode has inherent immunity, defensive capabilities, and adaptive capacities against disruption events, bringing opportunities for resilience governance in complex product R&D networks.
In complex product R&D network, focal firms’ dominant interactive behaviors are common strategic activities under disruption contexts. Firms generally believe that focal firm-dominated mode enable firms targeted acquisition of specialized information and resources [22]. For focal firm, this ensures stable input of effective resources, improves resource orchestration efficiency, thereby enhancing self-organized recovery capabilities during disruptions and accelerating response to disruption events. In practice, the focal firm-dominated mode, which is self-development-oriented, breaks the basic principle of collaborative symbiosis in complex product R&D networks [23], undermining network resilience governance. Specifically, focal firms engage in asymmetric interactions with other network participants through their central positions in complex product R&D networks, resulting in significant differences in relational strength indicators including contact frequency, collaboration scope, intimacy, and trust between the focal and other entities [24]. Such relational strength changes mean the inherent network structure and inertial cooperation mechanisms obsolete. The disruption events further strengthen the negative effects of asymmetric interactions under the focal firm-dominated mode. Thus, the focal firm-dominated mode helps focal firms to quickly adapt to the disruptions, but this context-dependent strategy sacrifices collaborative bonds among complex product R&D network multi-agents, representing short-sighted governance that aggravates network vulnerability and impedes long-term resilience enhancement. Therefore, under disruption events, multi-agent co-creation mode demonstrate superior effectiveness for complex product R&D network resilience governance compared to focal firm-dominated mode.
Thus, we propose the following hypotheses:
H3. 
The multi-agent co-creation mode effectively addresses disruption events and enhances complex product R&D networks resilience.
H4. 
The focal firm-dominated mode is not conducive to enhancement of the complex product R&D networks resilience during disruption events.
In sum, the study constructs the theoretical model shown in Figure 1.

3. Research Method

3.1. Sample Selection and Data Collection

The study focuses on core enterprises in the high-end equipment manufacturing within China’s Yangtze River Delta and Pearl River Delta regions. The reason for this selection is multifaceted: First, in terms of industry characteristics. High-end equipment manufacturing products align with the characteristics of complex products, such as high costs, advanced technological integration, and small-batch or customized production [25]. Second, industrial context. The development of high-end equipment manufacturing products involves multi-agent collaboration, joint R&D participation, and cross-team technology integration. Third, regional advantages. The Yangtze River Delta and Pearl River Delta regions provide a fertile ecosystem for high-end equipment manufacturing, fostering industrial clusters and nurturing numerous leading enterprises in this sector. Fourth, research team accessibility. A significant portion of the research team is based in the Yangtze River Delta and Pearl River Delta regions, granting significant geographical advantages in organizing field investigations and conducting in-depth interviews. In summary, selecting high-end equipment manufacturing enterprises from Yangtze River Delta and Pearl River Delta regions combines typicality and representativeness as research subjects with accessibility and convenience in data acquisition. This aligns with the study’s contextual requirements for studying multi-agent interactions in complex product R&D networks.
Supported by the Guangdong Shunde Rural Commercial Bank and the Jiangsu Rural Commercial United Bank, the study collects data through questionnaires distributed to core enterprises in high-end equipment manufacturing across the Yangtze River Delta and Pearl River Delta regions from March to December 2024. The questionnaire distribution employed two methods: First, face-to-face interviews. We interviewed 3 credit clients of the Shunde Rural Commercial Bank and 6 credit clients of the Jiangsu Rural Commercial United Bank, yielding 9 valid questionnaires. Second, snowball sampling and online outreach. Using the above 9 enterprises as the center, we distributed 859 questionnaires via snowball sampling and network communication tools. Of these, 582 responses were collected. After removing 164 invalid responses (due to patterned answers or missing critical items), 418 valid questionnaires were retained, with an effective response rate of 48.66%. The response rate is comparable to those reported in existing resilience studies, and falls within the standard range for survey research, thereby meeting the norms of social science research. Combining both methods, a total of 427 valid questionnaires were ultimately obtained. The basic information of the sample enterprises is summarized in Table 1.

3.2. Measurements

All the items were measured on a 5-point Likert scale (1 = “Strongly Disagree”; 5 = “Strongly Agree”). Appendix A presents the measures we used in the study.
Disruption Events (DE). Disruption events were measured with the six-item scale from Anderson and Lewis [10].
Multi-Agent Interaction Modes. We conceptualized multi-agent interaction modes as a multidimensional construct comprising two distinct but related dimensions: focal firm-dominated mode (FELM) and multi-agent co-creation mode (MACM). We adapted a three-item measurement scale from Gao [6] to measure focal firm-dominated mode, and multi-agent co-creation mode was measured with 13 items developed by Ren et al. [26].
Complex Product R&D Network Resilience (NR). Nine items were adapted from Liang et al. [27] to measure complex product R&D network resilience.
Control Variables. Following the recommendation of previous studies [28], we selected the following variables as controls. Since firms with greater experience and larger scale are more likely to possess stronger network relationships, we controlled for firm age (Age) and size (Size). The ownership advantages of state-owned enterprises enable them to gain government support more easily. Accordingly, state-owned or state-controlled firms tend to be more stable. Thus, we controlled for ownership type (Ownership). Finally, although both the Yangtze River Delta and the Pearl River Delta represent developed economic regions, the mechanisms underlying network formation differ substantially: networks in the Yangtze River Delta are primarily based on business collaboration, whereas those in the Pearl River Delta are significantly influenced by clan culture. Due to the differences in network formation, we controlled for geographical location (Location).

3.3. Reliability, Validity, and Common Method Variance

The reliability of these constructs was evaluated with Cronbach’s alphas and composite reliability. As shown in Table 2, Cronbach’s alphas range from 0.808 to 0.943, which are above 0.70 threshold. We also found that the values of CRs, ranging from 0.887 to 0.953, were above the recommended level, suggesting satisfactory reliability.
We assessed the discriminant validity of these constructs. As shown in Table 2, the average variance extracted (AVE) exceeded the relevant squared correlation, providing evidence of discriminant validity.
Next, we further assessed the factor structure and validity of the latent variables in our model through a confirmatory factor analysis (CFA). As shown in Table 3, the hypothesized four-factor model fit the data better when compared to other alternative models (χ2/df = 2.636, NFI = 0.858, TLI = 0.899, CFI = 0.907, RMSEA = 0.062). All items loaded substantially, were statistically significant on their latent factor, and all factor loadings were above 0.724, indicating satisfactory convergent validity.
In addition, we employed a few more techniques to address common method bias. First, Harman’s single-factor test indicated that no single factor accounted for the majority of the variance (the first factor accounted for 25.026% of the 67.071% explained variance). Second, compared to the measurement model, the single-factor model had a significantly poorer fit to data (χ2/df = 10.591, NFI = 0.423, TLI = 0.405, CFI = 0.445, RMSEA = 0.150). Third, we used the marker variable method. Following Chin et al.’s [29] steps for common method bias testing, the test results showed that the marker variable did not have a significant impact on the results of all research hypotheses. Hence, we believed that common method bias did not seem to be a serious concern in the research.

4. Empirical Analysis

Table 4 presents the regression results. Model 1 uses the focal firm-dominated mode as the dependent variable, Model 2 uses the multi-agent co-creation mode, Model 3 uses the ratio of the focal firm-dominated mode to the multi-agent co-creation mode, and Models 4–7 use the resilience of complex product R&D networks as the dependent variable.

4.1. Disruption Events and Multi-Agent Interaction Modes in Complex Product R&D Networks

The results of Model 1 in Table 4 show a significant positive correlation between disruption events and the focal firm-dominated mode. Model 2 shows that disruption events are significantly positively correlated with the multi-agent co-creation mode. These results indicate that complex product R&D networks under disruption events are prone to two multi-agent interaction modes: the focal firm-dominated mode and the multi-agent co-creation mode. Thus, H1 is validated.
To examine the prioritization of attention allocation between the focal firm-dominated mode and the multi-agent co-creation mode in the complex product R&D network under disruption events, we adopted Fan et al.’s [30] approach. First, we construct a ratio variable of the focal firm-dominated mode to the multi-agent co-creation mode. Second, the ratio is used as the dependent variable, with disruption events as the independent variable, to fit the model. The ratio eliminates commonalities between the two multi-agent interaction modes while highlighting their differences, thereby reflecting the portion of the focal firm-dominated mode that exceeds the multi-agent co-creation mode. The results of Model 3 show that disruption events have a significantly positive impact on the ratio of the focal firm-dominated mode to the multi-agent co-creation mode, suggesting that disruption events exert a stronger influence on the focal firm-dominated mode than on the multi-agent co-creation mode. In other words, compared to the multi-agent co-creation mode, focal firms in complex product R&D networks under disruption events prioritize allocating attention to the focal firm-dominated mode. Thus, H2 is validated.

4.2. Multi-Agent Interaction Modes and Resilience of Complex Product R&D Networks

Model 4 in Table 4 tested the direct effect of disruption events on complex product R&D network resilience. The results showed that complex product R&D networks can enhance resilience through correct multi-agent interaction mode selection under disruption events. Model 5 included both disruption events and the focal firm-dominated mode. The results show a significantly positive correlation between the focal firm-dominated mode and network resilience, which remains robust in Model 7. This indicates that the focal firm-dominated mode helps complex product R&D networks cope with disruptions and improve resilience, contradicting H4. Model 6 included both disruption events and the multi-agent co-creation mode. The results show a significantly positive correlation between the multi-agent co-creation mode and network resilience, which also remains robust in Model 7. This suggests that the multi-agent co-creation mode is an effective strategy for building resilient complex product R&D networks under disruption events. H3 is validated.
Given that both the focal firm-dominated mode and the multi-agent co-creation mode positively influence the complex product R&D network resilience, we compare which multi-agent interaction mode plays a greater role in enhancing resilience under disruption events. Specifically, ΔR2(M7 – M5) = R2(M7) − R2(M5) = 0.065 − 0.053 = 0.012, while ΔR2(M7 − M6) = R2(M7) − R2(M6) = 0.065 − 0.059 = 0.006. The former reflects the variance in network resilience explained by the multi-agent co-creation mode, whereas the latter reflects the variance explained by the focal firm-dominated mode. The result ΔR2(M7 − M5) > ΔR2(M7 − M6) indicates that, under disruption events, the multi-agent co-creation mode accounts for a greater degree of improvement in the complex product R&D network resilience compared with the focal firm-dominated mode.

4.3. Further Research

First-order and second-order disruption events differ in both their sources of occurrence and their signaling characteristics. Thus, we refine the disruption events and further explore the effects of multi-agent interaction modes under different level of disruptions. First-order disruptions impose slight impacts on complex product R&D networks. The focal firm-dominated mode helps the complex product R&D network to repair the network interaction system in a short time and ensure business continuity, while the multi-agent co-creation mode provides a guarantee for the systematic and sustainable development of the complex product R&D network. Second-order disruptions lead to a serious impact on the complex product R&D network, where collective collaboration and co-creation among multiple agents are critical to maintaining stable network structures and orderly interactions [31,32]. The focal firm-dominated mode, which prioritizes self-interest, risks disrupting network relationships. Thus, we hypothesize that under first-order disruptions, both the focal firm-dominated mode and the multi-agent co-creation mode enhance complex product R&D network resilience, whereas under second-order disruptions, only the multi-agent co-creation mode contributes to resilience.
To verify the above hypothesis, we constructed two extended indicators based on the disruption event scale: the first three items measure first-order disruptions with a Cronbach’s alpha of 0.814, while the latter three items measure second-order disruptions with a Cronbach’s alpha of 0.741. Table 5 presents the effects of multi-agent interaction modes on the resilience of complex product R&D networks under different levels of disruption events. Model 1 includes both first-order disruption events and the focal firm-dominated mode, whereas Model 2 includes first-order disruption events and the multi-agent co-creation mode. The results show that both the focal firm-dominated mode and the multi-agent co-creation mode exhibit significant positive correlations with the resilience of complex product R&D networks, and these findings remain robust in Model 3. This indicates that multi-agent interaction is an effective mechanism for complex product R&D networks to address the first-order disruption events. Furthermore, we compare the relative efficacy of the focal firm-dominated mode versus the multi-agent co-creation mode in enhancing complex product R&D network resilience under first-order disruptions. ΔR2(M3 − M1) = R2(M3) − R2(M1) = 0.060 − 0.046 = 0.014, ΔR2(M3 − M2) = R2(M3) − R2(M2) = 0.060 − 0.052 = 0.008, the former represents the resilience variance proportion of complex product R&D network explained by the multi-agent co-creation mode, and the latter represents the variance proportion explained by the focal firm-dominated mode. The inequality ΔR2(M3 − M1) > ΔR2(M3 − M2) implies that, under first-order disruptions, the multi-agent co-creation mode accounts for a greater improvement in complex product R&D network resilience compared to the focal firm-dominated mode. Model 4 includes second-order disruption events and the focal firm-dominated mode, while Model 5 includes second-order disruption events and the multi-agent co-creation mode. The results show no significant correlation between the focal firm-dominated mode and complex product R&D network resilience, whereas the multi-agent co-creation mode maintains a significant positive correlation with complex product R&D network resilience. These findings remain robust in Model 6, suggesting that only the multi-agent co-creation mode helps the complex product R&D network to cope with second-order disruption events. Overall, these results not only validate the hypothesis proposed in the further analysis, but also corroborate the original theoretical proposition: for complex product R&D networks to effectively deal with the disruption events, the multi-agent co-creation mode is superior to the focal firm-dominated mode.

5. Discussion

Through a series of analyses, we find that disruption events drive multi-agent interactions in complex product R&D networks, giving rise to two modes: the focal firm-dominated mode and the multi-agent co-creation mode. Further analysis of differentiated disruption events shows that under first-order disruptions, both the focal firm-dominated mode and the multi-agent co-creation mode are significantly and positively correlated with the complex product R&D network resilience. Moreover, compared with the focal firm-dominated mode, the multi-agent co-creation mode explains a greater degree of improvement in network resilience. Under second-order disruptions, however, the focal firm-dominated mode shows no significant relationship with complex product R&D network resilience, while the multi-agent co-creation mode remains significantly and positively correlated with it.

5.1. Theoretical Contributions

First, the existing literature has extensively studied network resilience in the context of major disruption incidents such as natural disasters, geopolitical conflicts, and public health crises [33,34], generating rich theoretical insights. However, unlike these sudden shocks, the disruption events analyzed in the study encompass not only infrequent, high-order, and highly destructive external events (i.e., second-order disruptions), but also frequent, low-order, and less destructive internal events (i.e., first-order disruptions). Therefore, investigating network resilience by considering both a comprehensive perspective of disruptions and a nuanced perspective of disruptions constitutes a marginal theoretical contribution.
Second, although prior studies have explored how network structures composed of limited participants affect product development resilience [35], actors in complex product R&D projects are far more diverse. This broader participation highlights the limitations of an interorganizational interaction perspective and underscores the necessity of introducing a multi-agent interaction perspective. Building on this perspective, the study examines two modes—focal firm-dominated and multi-agent co-creation—and investigates their effects on complex product R&D network resilience under disruptions. This approach not only helps to reconceptualize the micro-dynamics and diversity of multi-agent interactions in complex product systems but also offers a theoretical foundation for focal firms to govern complex product R&D network resilience by leveraging the characteristics of multi-agent interaction.
Third, although some scholars have recognized the importance of firms’ attentional resources during disruptions [36], they have largely overlooked the issue of how these resources are allocated. The study demonstrates, from the dual dimensions of focal firm-dominated and multi-agent co-creation modes, that responses to disruption events in complex product R&D networks are both complex and diverse, and that focal firms exhibit selectivity and preference in allocating attentional resources across different modes. This not only clarifies the mechanisms through which focal firms respond to disruptions but also extends the research framework of sequential attention logic.
Fourth, existing studies on disruption events often focused on firms’ response strategies, with little attention to whether these strategies achieve the intended outcomes. As a result, the research chain of disruption–response–recovery remains incomplete. By comparatively analyzing the effects of focal firm-dominated and multi-agent co-creation modes on resilience governance under disruptions, the study finds that multi-agent co-creation constitutes a more effective strategy for complex product R&D networks to deal with disruption events. This finding not only addresses the challenge of selecting among multi-agent interaction modes under disruptions but also remedies the lack of attention to the recovery stage in the existing disruption research chain.

5.2. Practical Implications

First, although disruption events negatively impact interorganizational collaboration, they may also provide opportunities for optimizing network governance. In complex product R&D networks that emphasize multi-agent collaboration, focal firms should strategically leverage disruptions to strengthen interactive relationships with upstream and downstream supply chain partners, thereby achieving symbiotic coexistence among multiple actors. Second, under disruption events, complex product R&D networks tend to exhibit two multi-agent interaction modes. Focal firms may prioritize self-interest and adopt a focal firm-dominated mode, or they may emphasize collective interests and engage in a multi-agent co-creation mode. It is worth noting that in the context of disruptions, focal firms often seek rapid normalization and tend to focus on their own development and short-term benefits, which makes them more inclined toward a focal firm-dominated mode. While this mode can be effective in responding to first-order disruptions, it fails to generate the expected outcomes when confronted with second-order disruptions. In VUCA environments, multi-agent collaboration of complex product R&D networks and the rhythm of complex product R&D are highly vulnerable to disruption shocks. Building resilience complex product R&D network requires balancing collective and individual interests. Only by engaging diverse network actors in interactive co-creation and enhancing collaborative synergies can focal firms strengthen the resilience of complex product R&D networks and ensure the stable functioning of complex product R&D activities. Third, government should reinforce the construction of industrial ecosystems by promoting shared value propositions and facilitating broader and deeper collaborative mechanisms, thereby supporting resilience governance in complex product R&D network.

6. Conclusions

This study yields three key findings: ① Under disruption events, complex product R&D networks are prone to two multi-agent interaction modes: the focal firm-dominated mode and the multi-agent co-creation mode. ② Compared to the multi-agent co-creation mode, the focal firm-dominated mode is more likely to emerge in complex product R&D networks under disruption events. ③ Under first-order disruptions, both the focal firm-dominated mode and the multi-agent co-creation mode are important strategies for resilience governance of complex product R&D networks. However, under second-order disruptions, only the multi-agent co-creation mode is a key mechanism for resilience governance.
Of course, the study has several limitations, which offers interesting opportunities for future research. First, while we validated the coexistence of different multi-agent interaction modes in complex product R&D networks under disruption events, we did not address the internal dynamic evolution among these modes [37]. Future research could explore the evolutionary logic of multi-agent interaction modes in complex product R&D networks from perspectives such as lifecycle dynamics. Second, we analyze the microeconomic effects of disruption events from a static perspective, overlooking their time-varying characteristics [38]. Subsequent studies may integrate temporal elements into the research framework to construct a time-varying model for multi-agent interaction modes in complex product R&D networks under disruption-recovery mechanisms. Third, we employ the high-end equipment manufacturing industry network as a proxy for complex product R&D networks. However, this industry includes multiple sub-sectors (e.g., manufacturing equipment, engineering equipment). Future research could extend the investigation to these specialized sub-industries to enhance generalizability.

Author Contributions

Conceptualization, S.C. and Q.F.; methodology, S.C.; resources, Q.F.; writing—original draft preparation, S.C.; writing—review and editing, Q.F. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the Humanities and Social Sciences Research Project of the Ministry of Education, grant number 23YJCZH026, and the ‘14th Five-Year Plan’ of Philosophy and Social Sciences in Guangzhou, grant number 2025GZGJ289.

Data Availability Statement

Dataset available on request from the authors. The raw data supporting the conclusions of this article will be made available by the authors on request.

Conflicts of Interest

The authors declare no conflicts of interest.

Appendix A

Measures of key variables
Disruption Events (Anderson and Lewis [10]) (Anderson and Lewis, 2014)
1. Employee turnover occurs from time to time within the company.
2. The company frequently faces challenges arising from technological change.
3. The company often encounters shifts in task assignments.
4. The company experiences changes in market structure.
5. The company is confronted with transformations in industry business models.
6. Extreme and unpredictable events occasionally occur in the company.
Multi-Agent Interaction Modes—Focal Firm-Dominated Mode (Gao [6])
1. The company proactively designs and plans the cooperative framework and rules of its R&D network.
2. The company takes the lead in guiding and adjusting interactions with partners based on its own development objectives.
3. The company deliberately prioritizes the acquisition and integration of critical resources within the network.
Multi-Agent Interaction Modes—Multi-Agent Co-Creation Mode (Ren et al. [26])
1. The company and its R&D partners exchange information frequently, either irregularly or on a scheduled basis.
2. The company seeks to maintain active communication with its R&D partners whenever possible.
3. Communication between the company and its partners is open and constructive.
4. When disagreements arise, the company and its partners resolve them through dialogue.
5. Partners can easily access information regarding the company’s product development from multiple sources.
6. Partners are not required to obtain ownership of the company’s products in order to access product development information.
7. When existing product development falls short of company needs, partners help the company acquire desired R&D information through new channels.
8. The company invites partners to jointly evaluate and share risks.
9. Partners agree to co-create value and are therefore willing to share R&D risks with the company.
10. The company establishes dedicated mechanisms for risk assessment and avoidance to help both itself and partners mitigate common R&D risks.
11. The company maintains high transparency toward its partners.
12. The company treats its partners with sincerity and does not conceal critical R&D information.
13. Potential risks to partners caused by information asymmetry with the company are minimal.
Complex Product R&D Network Resilience (Liang et al. [27])
1. The company and its partners are able to jointly forecast risks associated with various disruptions.
2. The company and its partners are able to jointly develop contingency plans for different types of disruptions.
3. The company and its partners regularly assess the external environment together to identify potential disruption signals in a timely manner.
4. In the event of disruptions, the company and its partners can still ensure the continuity of product development.
5. In the event of disruptions, the company and its partners can rely on multiple backup plans to sustain R&D activities.
6. When R&D interruptions occur, the company and its partners can quickly mobilize resources or relationships to maintain normal operations.
7. Employees of the company and its partners are proactive in identifying and learning new knowledge and skills.
8. After experiencing disruptions, the company and its partners jointly reflect and improve existing collaboration processes and response mechanisms.
9. When facing unprecedented challenges, the company and its partners can rapidly build new capabilities or develop innovative solutions together.

References

  1. Hobday, M.; Rush, H.; Tidd, J. Innovation in complex products and system. Res. Policy 2000, 29, 793–804. [Google Scholar] [CrossRef]
  2. Zuzul, T.W. “Matter battles”: Cognitive representations, boundary objects, and the failure of collaboration in two smart cities. Acad. Manag. J. 2019, 62, 739–764. [Google Scholar] [CrossRef]
  3. Yang, N.D.; Wang, Y.; Wang, Y. Evolutionary game on technological innovation of enterprises in complex product R&D network. Chin. J. Manag. Sci. 2024, 32, 242–253. [Google Scholar]
  4. Ruan, F.; Feng, N.; Wei, F.; Zhao, R.; Yang, S. How do risks affect innovation performance in CoPS projects? A configurational perspective based on a technology–organization–environment framework. Manag. Decis. 2025. [Google Scholar] [CrossRef]
  5. Randolph, R.V.G. A multilevel study of structural resilience in interfirm collaboration: A network governance approach. Manag. Decis. 2016, 54, 248–266. [Google Scholar] [CrossRef]
  6. Gao, X. The Mechanism of the Mode of Multi-Agent Interaction to Opportunity Realization of Digital Firms; Jilin University: Changchun, China, 2024. [Google Scholar]
  7. Ocasio, W. Towards an attention-based view of the firm. Strateg. Manag. J. 1997, 18 (Suppl. S1), 187–206. [Google Scholar] [CrossRef]
  8. Joseph, J.; Laureiro-Martinez, D.; Nigam, A.; Ocasio, W.; Rerup, C. Research Frontiers on the Attention-Based View of the Firm; SAGE Publications: London, UK, 2024; Volume 22, pp. 6–17. [Google Scholar]
  9. Sullivan, B.N. Competition and beyond: Problems and attention allocation in the organizational rulemaking process. Organ. Sci. 2010, 21, 432–450. [Google Scholar] [CrossRef]
  10. Anderson, E.G.; Lewis, K. A dynamic model of individual and collective learning amid disruption. Organ. Sci. 2014, 25, 356–376. [Google Scholar] [CrossRef]
  11. Brielmaier, C.; Friesl, M. The attention-based view: Review and conceptual extension towards situated attention. Int. J. Manag. Rev. 2023, 25, 99–129. [Google Scholar] [CrossRef]
  12. Joseph, J.; Wilson, A.J. The growth of the firm: An attention-based view. Strateg. Manag. J. 2018, 39, 1779–1800. [Google Scholar] [CrossRef]
  13. Greve, H.R. A Behavioral Theory of Firm Growth: Sequential Attention to Size and Performance Goals. Acad. Manag. J. 2008, 51, 476–494. [Google Scholar] [CrossRef]
  14. Kumar, P.; Zaheer, A. Ego-network stability and innovation in alliances. Acad. Manag. J. 2019, 62, 691–716. [Google Scholar] [CrossRef]
  15. Kim, Y.; Chen, Y.-S.; Linderman, K. Supply network disruption and resilience: A network structural perspective. J. Oper. Manag. 2015, 33, 43–59. [Google Scholar] [CrossRef]
  16. Keil, T.; Posen, H.E.; Workiewicz, M. Aspirations, beliefs and a new idea: Building on march’s other model of performance feedback. Acad. Manag. Rev. 2023, 48, 749–771. [Google Scholar] [CrossRef]
  17. Cheng, S.; Fan, Q.; Song, Y. Performance gap and innovation ambidexterity: A moderated mediation model. Sustainability 2023, 15, 3994. [Google Scholar] [CrossRef]
  18. Storbacka, K.; Brodie, R.J.; Böhmann, T.; Maglio, P.P.; Nenonen, S. Actor engagement as a microfoundation for value co-creation. J. Bus. Res. 2016, 69, 3008–3017. [Google Scholar] [CrossRef]
  19. Chowdhury, I.N.; Gruber, T.; Zolkiewski, J. Every cloud has a silver lining—Exploring the dark side of value co-creation in B2B service networks. Ind. Mark. Manag. 2016, 55, 97–109. [Google Scholar] [CrossRef]
  20. Naug, D. Structure and resilience of the social network in an insect colony as a function of colony size. Behav. Ecol. Sociobiol. 2009, 63, 1023–1028. [Google Scholar] [CrossRef]
  21. Gorelova, I.; Dmitrieva, D.; Dedova, M.; Savastano, M. Antecedents and consequences of digital entrepreneurial ecosystems in the interaction process with smart city development. Adm. Sci. Q. 2021, 11, 94. [Google Scholar] [CrossRef]
  22. Kim, D.Y.; Zhu, P. Supplier dependence and R&D intensity: The moderating role of network centrality and interconnectedness. J. Oper. Manag. 2018, 64, 7–18. [Google Scholar] [CrossRef]
  23. Gastaldi, L.; Appio, F.P.; Trabucchi, D.; Buganza, T.; Corso, M. From mutualism to commensalism: Assessing the evolving relationship between complementors and digital platforms. Inf. Syst. J. 2024, 34, 1217–1263. [Google Scholar]
  24. Jiafu, S.; Yu, Y.; Tao, Y. Measuring knowledge diffusion efficiency in R&D networks. Knowl. Manag. Res. Pract. 2018, 16, 208–219. [Google Scholar]
  25. Li, B.; Su, J.; Yuan, B.; Li, L.; Zhao, Y.; Qin, Z.; Qian, L. A supplier selection decision-making approach for complex product development based on hesitant fuzzy information. Axioms 2023, 12, 1006. [Google Scholar] [CrossRef]
  26. Ren, S.J.; Hu, C.; Ngai, E.; Zhou, M. An empirical analysis of inter-organisational value co-creation in a supply chain: A process perspective. Prod. Plan. Control 2015, 26, 969–980. [Google Scholar]
  27. Liang, L.; Zhao, Y.B.; Liu, B. Research on monitoring and early warning of the resilience of the innovation ecosystem in the China’s national new districts. China Soft Sci. Mag. 2020, 51, 92–111. [Google Scholar]
  28. Do, H.; Budhwar, P.; Shipton, H.; Nguyen, H.-D.; Nguyen, B. Building organizational resilience, innovation through resource-based management initiatives, organizational learning and environmental dynamism. J. Bus. Res. 2022, 141, 808–821. [Google Scholar] [CrossRef]
  29. Chin, W.W.; Thatcher, J.B.; Wright, R.T.; Steel, D. Controlling for Common Method Variance in PLS Analysis: The Measured Latent Marker Variable Approach, New Perspectives in Partial Least Squares and Related Methods; Springer: Berlin/Heidelberg, Germany, 2013; pp. 231–239. [Google Scholar]
  30. Qunpeng, F.; Songsong, C.; Jia, L. Opening the black box between performance gaps and business model innovation: The effect of network strategy. Technol. Anal. Strateg. Manag. 2024. [Google Scholar] [CrossRef]
  31. Azara, I.; Pappas, N.; Michopoulou, E. Revisiting value cocreation and codestruction in events: An overview. Event Manag. 2023, 27, 157–162. [Google Scholar] [CrossRef]
  32. Assiouras, I.; Vallström, N.; Skourtis, G.; Buhalis, D. Exploring how interaction leads to value co-creation, value co-destruction, and value no-creation during service mega-disruptions. Curr. Issues Tour. 2024, 27, 3231–3248. [Google Scholar]
  33. El Nayal, O.; Slangen, A.; van Oosterhout, J.; van Essen, M. Towards a democratic new normal? Investor reactions to interim-regime dominance during violent events. J. Manag. Stud. 2020, 57, 505–536. [Google Scholar]
  34. Dutta, S. Creating in the crucibles of nature’s fury: Associational diversity and local social entrepreneurship after natural disasters in California, 1991–2010. Adm. Sci. Q. 2017, 62, 443–483. [Google Scholar] [CrossRef]
  35. Massari, G.F.; Giannoccaro, I.; Carbone, G. Team social network structure and resilience: A complex system approach. IEEE Trans. Eng. Manag. 2021, 70, 209–219. [Google Scholar] [CrossRef]
  36. Vanderkooi, D.; Mashatan, A.; Turetken, O. Introducing technological disruption: How breaking media attention on corporate events impacts online sentiment. J. Bus. Anal. 2024, 7, 63–82. [Google Scholar] [CrossRef]
  37. Shi, J.-L.; Chen, G.-H. Orchestrating multi-agent knowledge ecosystems: The role of makerspaces. Front. Psychol. 2022, 13, 898134. [Google Scholar] [CrossRef] [PubMed]
  38. Kilkki, K.; Mäntylä, M.; Karhu, K.; Hämmäinen, H.; Ailisto, H. A disruption framework. Technol. Forecast. Soc. Change 2018, 129, 275–284. [Google Scholar] [CrossRef]
Systems 13 00836 g001
Figure 1. Theoretical Model.
Figure 1. Theoretical Model.
Systems 13 00836 g001
Table 1. Descriptive statistics of sample enterprises.
Table 1. Descriptive statistics of sample enterprises.
Sample
Characteristics		CategoryFrequency%Sample
Characteristics		CategoryFrequency%
Age≤5 years10424.356Size≤500 employees8118.970
6–10 years11426.698501–1000
employees		10123.653
11–15 years429.8361001–1500
employees		18743.794
≥16 years16739.1101501–2000
employees		4510.539
Ownership
type		State-owned
enterprise		10023.419≥2000
employees		133.044
Nonstate-owned
enterprises		32776.581LocationYangtze River Delta19345.199
——Pearl River Delta23454.801
Table 2. Descriptive statistics, correlations, reliability and validity.
Table 2. Descriptive statistics, correlations, reliability and validity.
VariableMeanSDCronbach’s αCRAVECorrelations
1234
1. DE3.7280.9360.8770.9070.6211
2. FELM3.6811.0070.8080.8870.7240.287 **1
3. MACM4.0530.7990.9430.9500.5950.0840.0331
4. NR3.6610.7640.8540.9530.6950.150 **0.124 *0.126 **1
Notes: * Significant at 5%, ** Significant at 1%.
Table 3. Comparison of alternative factor structures.
Table 3. Comparison of alternative factor structures.
Modelχ2/dfNFITLICFIRMSEA
Four-factor model2.6360.8580.8990.9070.062
Three-factor model3.6640.8020.8350.8470.079
Two-factor model6.3290.6560.6700.6920.112
One-factor model10.5910.4230.4050.4450.150
Notes: One-factor model: DE+FELM+MACM+NR; Two-factor model: DE+FELM+MACM, NR; Three-factor model: DE, FELM+MACM, NR; Four factor model: Hypothesized Model.
Table 4. Disruption events and multi-agent interaction modes in complex product R&D networks.
Table 4. Disruption events and multi-agent interaction modes in complex product R&D networks.
VariableFELMMACMFELM/MACMNR
Model 1Model 2Model 3Model 4Model 5Model 6Model 7
Age0.040
(1.035)		0.024
(0.761)		0.005
(0.287)		−0.048
(−1.615)		−0.051 *
(−1.702)		−0.051 *
(−1.712)		−0.053 *
(−1.798)
Size0.024
(0.508)		−0.002
(−0.053)		0.013
(0.539)		−0.020
(−0.547)		−0.021
(−0.590)		−0.020
(−0.544)		−0.021
(−0.587)
Owner-ship−0.127
(−1.150)		−0.045
(−0.493)		−0.004
(0.072)		−0.201 **
(−2.340)		−0.193 **
(−2.247)		−0.196 **
(−2.295)		−0.188 **
(−2.202)
Location0.105
(1.113)		0.013
(0.173)		0.031
(0.649)		0.087
(1.184)		0.080
(1.093)		0.085
(1.170)		0.079
(1.080)
DE0.311 ***
(6.205)		0.073 *
(1.752)		0.058 **
(2.317)		0.124 ***
(3.180)		0.104 **
(2.561)		0.116 ***
(2.982)		0.096 **
(2.377)
FELM 0.064 *
(1.688)		 0.063 *
(1.683)
MACM 0.109 **
(2.398)		0.109 **
(2.393)
R20.0910.0090.0150.0460.0530.0590.065
F8.3940.7601.2414.0683.8804.3874.181
Note: * Significant at 10%, ** Significant at 5%, *** Significant at 1%.
Table 5. The influence of multi-agent interaction modes on the resilience of complex product R&D networks under different levels of disruption events.
Table 5. The influence of multi-agent interaction modes on the resilience of complex product R&D networks under different levels of disruption events.
VariableNR
Model 1Model 2Model 3Model 4Model 5Model 6
Age−0.051 *
(−1.727)		−0.051 *
(−1.730)		−0.054 *
(−1.824)		−0.048
(−1.625)		−0.048
(−1.643)		−0.050 *
(−1.718)
Size−0.019
(−0.533)		−0.017
(−0.474)		−0.019
(−0.534)		−0.023
(−0.635)		−0.021
(−0.595)		−0.022
(−0.630)
Owner-ship−0.198 **
(−2.291)		−0.204 **
(−2.370)		−0.193 **
(−2.244)		−0.180 **
(−2.111)		−0.181 **
(−2.136)		−0.176 **
(−2.078)
Location0.077
(1.054)		0.083
(1.134)		0.076
(1.043)		0.082
(1.135)		0.086
(1.200)		0.081
(1.121)
FDE0.069 *
(1.855)		0.085 **
(2.363)		0.065 *
(1.749)
SDE 0.150 ***
(3.859)		0.156 ***
(4.158)		0.141 ***
(3.624)
FELM0.072 *
(1.904)		 0.071 *
(1.875)		0.053
(1.418)		 0.053
(1.423)
MACM 0.115 **
(2.522)		0.114 **
(2.499)		 0.100 **
(2.219)		0.100 **
(2.221)
R20.0460.0520.0600.0710.0770.081
F3.3393.8133.7905.3235.8435.310
Note: * Significant at 10%, ** Significant at 5%, *** Significant at 1%.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Cheng, S.; Fan, Q. Research on the Selection of Multi-Agent Interaction Modes in Complex Product R&D Networks Under Disruption Events. Systems 2025, 13, 836. https://doi.org/10.3390/systems13100836

AMA Style

Cheng S, Fan Q. Research on the Selection of Multi-Agent Interaction Modes in Complex Product R&D Networks Under Disruption Events. Systems. 2025; 13(10):836. https://doi.org/10.3390/systems13100836

Chicago/Turabian Style

Cheng, Songsong, and Qunpeng Fan. 2025. "Research on the Selection of Multi-Agent Interaction Modes in Complex Product R&D Networks Under Disruption Events" Systems 13, no. 10: 836. https://doi.org/10.3390/systems13100836

APA Style

Cheng, S., & Fan, Q. (2025). Research on the Selection of Multi-Agent Interaction Modes in Complex Product R&D Networks Under Disruption Events. Systems, 13(10), 836. https://doi.org/10.3390/systems13100836

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop