Novel Multi-Criteria Decision Analysis Based on Performance Indicators for Urban Energy System Planning

Abstract

Urban energy systems planning presents significant challenges, requiring the integration of multiple objectives such as economic feasibility, technical reliability, and environmental sustainability. Although previous studies have focused on optimizing renewable energy systems, many lack comprehensive decision frameworks that address the complex trade-offs between these objectives in urban settings. Addressing these challenges, this study introduces a novel Multi-Criteria Decision Analysis (MCDA) framework tailored for the evaluation and prioritization of energy scenarios in urban contexts, with a specific application to the city of Bozen-Bolzano. The proposed framework integrates various performance indicators to provide a comprehensive assessment tool, enabling urban planners to make informed decisions that balance different strategic priorities. At the core of this framework is the Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS), which is employed to systematically rank energy scenarios based on their proximity to an ideal solution. This method allows for a clear, quantifiable comparison of diverse energy strategies, facilitating the identification of scenarios that best align with the city’s overall objectives. The flexibility of the MCDA framework, particularly through the adjustable criteria weights in TOPSIS, allows it to accommodate the shifting priorities of urban planners, whether they emphasize economic, environmental, or technical outcomes. The study’s findings underscore the importance of a holistic approach to energy planning, where trade-offs are inevitable but can be managed effectively through a structured decision-making process. Finally, the study addresses key gaps in the literature by providing a flexible and adaptable tool that can be replicated in different urban contexts to support the transition toward 100% renewable energy systems.

1. Introduction

Recent events have underscored the urgent need to transition to a sustainable energy system, including extreme weather phenomena [1,2], geopolitical tensions affecting fossil fuel supplies [3,4,5], energy market volatility [6,7], the increasing frequency of natural disasters [1,8], and rising pollution levels [9,10]. These events highlight the vulnerabilities of current energy systems from both technical and economic perspectives [11,12]. As a consequence, the ongoing energy transition needs an acceleration to buffer these pressing current issues and, thus, to shift towards renewable and resilient energy solutions [7,13]. However, achieving such an ambitious goal as a $100\%$ renewable energy system requires appropriate planning, taking into account the continuously changing socio-economic-environmental conditions and the evolution of various technologies [14,15]. In this respect, many aspects of the energy planning process need to be adapted to the new requirements, starting from energy modelling [16] to the optimization methodology [17] for the various objectives and finally to the decision-making process [18]. Thus, effective change toward a renewable-based future requires concerted efforts and actions from countries, cities, and individuals alike [19,20,21].

The $100\%$ renewable energy system at the global and regional scale represents the final goal and, as a consequence, exhibits various challenges and opportunities. Studies emphasize the importance of transitioning towards cleaner energy sources to reduce CO₂ emissions and combat climate change [22,23,24]. The integration of renewable energy sources like solar and wind power is crucial, with islands serving as ideal environments for showcasing technical solutions and transition pathways [25]. Furthermore, the role of energy storage technologies, such as batteries and hydrogen storage, is essential in managing fluctuations in renewable energy sources and electricity demand, contributing to the feasibility of a fully renewable energy system [26]. While there is a growing trend towards $100\%$ renewable energy systems, it is acknowledged that achieving this goal poses technical and economic challenges that require innovative solutions and comprehensive planning [27].

On a city scale, there have been several works to understand the technical and economic feasibility of transitioning to $100\%$ renewable energy in the urban environment, which highlights the importance of factors like energy job sector growth, land requirements, and investment recovery [28,29]. The potential benefits of this transition have been further stressed in these works, which include the reductions in primary energy consumption, cost and greenhouse gas emission while showing the inconsistencies in investment recovery and emission reductions across different renewable energy systems in urban settings [29]. There is also a need for robust policies, infrastructure development, and multi-governance approaches to accelerate the energy transition and achieve climate neutrality in cities by leveraging renewable energy sources [28,30]. Additionally, there are other efforts to achieve a $100\%$ renewable energy system in places like the residential community of Huanglong Township Island in Zhejiang province, China [31], the campus of Cornell University in the United States of America (USA) [32] and the local municipalities in Fukuoka, Japan [33].

This paper bridges the gab that exist in the energy system modelling phase and the decision making phase. It deals with the investigation of $100\%$ solutions at an urban scale for optimal planning through decision analysis; although it is the final step, it is a crucial stage in the urban energy design process supporting the sensitive process of the decision-making [34]. Decision analysis aims to analyze a set of solutions identified through rigorous studies, which typically include energy system modelling and the search for near-optimal configurations of scenarios [35] or pathways [36]. Multi-Criteria Decision Analysis (MCDA) evaluates a pool of near-optimal solutions, enabling a comprehensive analysis of trade-offs between various energy strategies. This analysis considers multiple evaluation objectives (e.g., technical, economic, and environmental) and different sources of uncertainty (e.g., climate scenarios and demographic trends) [37,38].

In energy systems planning, decision analysis plays a crucial role in managing the complexities of shifting to sustainable energy sources [39]. Among the various approaches, MCDA stands out for its ability to evaluate and balance multiple conflicting criteria [40], such as technical feasibility [41], economic viability [42], and environmental impact [43]. This approach is particularly beneficial when planning for renewable energy systems, where diverse factors have to be considered to develop effective and sustainable strategies [44]. MCDA allows planners to systematically assess different energy scenarios, weighing the pros and cons of each option. It provides a structured framework that helps decision-makers identify the best solutions from a pool of alternatives [45], considering not only immediate costs and benefits but also long-term implications [46]. This comprehensive evaluation is essential for ensuring that selected energy strategies are both practical and aligned with broader sustainability goals [47].

The Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) is especially useful within the MCDA framework. TOPSIS ranks alternative solutions based on their closeness to an ideal solution that maximizes desirable attributes and minimizes undesirable ones [48]. This method is advantageous because it offers clear, interpretable rankings of different alternatives, making it easier for decision-makers to make informed decisions [49]. By applying TOPSIS, we aim to provide energy planners with a tool capable of evaluating various energy scenarios in a systematic and transparent manner. This method helps to identify the most balanced and effective options, considering all relevant criteria [50]. As a result, TOPSIS supports robust, evidence-based decision-making, ensuring that the chosen solutions are technically sound, economically feasible, and environmentally sustainable [51]. This approach can be strategic for achieving the goal of effectively transitioning to renewable energy systems.

In this paper, the authors aim to seek the challenge of analyzing and interpreting the near-optimal solutions for urban energy systems planning in the case of multi-objective technical-economical-environmental optimization incorporating climate change. The main novelty lies in the proposed MCDA-based methodology, which uses performance indicators to flexibly analyze various conflicting aspects of energy planning (e.g., the cost of investing in and implementing state-of-the-art technologies, the use of renewable resources such as biomass and ensuring their sustainable use, and the environmental impact of the renewable energy system configurations) while also providing robust solutions to support decision-making processes. The proposed methodology is developed based on an urban case study of the alpine city of Bozen-Bolzano, where a pool of near-optimal configurations of the energy system according to technical, economic and environmental has been studied in [52,53]. Based on these solutions, the authors implemented the MCDA using a series of performance indicators covering different technical-economic-environmental aspects of the solutions considered. The indicators involved covered critical aspects of the design of urban renewable-based energy systems, such as Mismatch Compensation Factor, Emissions Reduction Effectiveness, Biomass System Efficiency, and Curtailment Fraction, among others. Instead, the MCDA proposes a comprehensive and robust investigation of the best solutions with respect to different types of targets.

The remainder of the paper details a case study of the energy system in Bozen-Bolzano, describing the materials and methodology, which include performance indicators and multi-criteria decision analysis in Section 2. The study’s results are presented in Section 3, followed by the final remarks in Section 4.

2. Case Study, Materials and Methods

2.1. Case Study

The Municipality of Bozen-Bolzano is the case study employed for developing and testing the proposed methodology in Section 2.3. Bozen-Bolzano is a city found in the North-Eastern part of Italy, in the centre of the Alps, with a population of 106,000. The climate is characterized by cold winters and hot summers, with 2328 °C Heating Degree Days (HDD) and 222 °C Cooling Degree Days (CDD)for the typical year [52,54,55]. This city offers distinct opportunities and challenges for the energy transition due to unique geographical and extreme climatic conditions [56].

The actual energy system of Bozen-Bolzano city is characterized by a transition towards smart energy city (SEC) concepts spearheaded by the integration of renewable energy sources, which includes photovoltaic and thermal solar panels, a growing district heating system fed mainly by a waste-to-energy plant, hydropower storage and run of the river plants, which aims to accelerate decarbonization efforts [57]. The city’s energy system emphasizes the use of renewable energy sources to improve overall energy efficiency and reduce its environmental impacts [58]. By combining various renewable energy sources, energy storage devices and advanced technologies, Bozen-Bolzano is striving towards sustainable economic development and energy independence, setting a modern example for other urban cities to follow [57,59]. More details are reported in [52].

2.2. Background Information

The background information forming the basis of the proposed decision analysis is derived from two previous studies. The final outcomes of these studies, which serve as inputs for the current decision analysis methodology, consist of a set of near-optimal solutions representing different configurations of $100\%$ renewable-based energy systems. Menapace et al. (2020) [52] address the first steps in designing a $100\%$ renewable energy system in Bozen-Bolzano by 2050. They first proposed a path by focusing on the operation between the sustainable use of biomass, replacement of traditional flexibility residing in fossil fuels with modern ones based on smart energy systems, balancing import and export of electricity and management of exchange of energy between the local energy system and it surrounding systems. Their approach integrated the energy system modelling using EnergyPLAN (a tool designed by the Sustainable Energy Planning Research Group at the Aalborg University to support the analysis of complex energy system factoring in advance technologies [60,61,62]) with multi simulations to accurately investigate the best technical alternative to achieve the main aim of a $100\%$ renewable energy system.

Building on this work, Battini et al. (2024) [53] further integrated energy system design with the impacts of climate change on energy demand and renewable production. Then, the optimal scenarios are searched by varying the installed capacity of Photo Voltaics (PV), Combined Heat and Power (CHP) and Heat Pumps (HP) through a multi-objective optimization approach considering economic, environmental and technical targets. Two methods were adopted for the best scenarios identification: Grid Search and Non-dominated Sorting Genetic Algorithm-II (NSGA2). For the purposes of this work, we consider the scenarios resulting from the grid search as near-optimal due to their performance comparable to NSGA2 but in a regular grid.

2.3. Decision Analysis Methodology

The overall energy planning strategy is described in Figure 1 from the beginning with the data collection and the modelling of the actual energy system to the final best scenarios identification. Specifically, steps 1–3 regard the energy modelling of the system, including the actual year and climate projections, the optimization and selection of the near-optimal scenarios [52,53]. The next steps 4–6 belong to the proposed work and include the definition of the performance indicator for a wide-ranging evaluation of the different scenarios and a comparison analysis of the best scenario through a new MCDA methodology.

Thus, this decision analysis includes these two main phases that merge the robustness of the performance indicator and the flexibility of MCDA. Thus, specifically, the former phase consists of the evaluation of the selected performance indicators that reflect the key targets of the $100\%$ renewable energy system: environmental impact, economic feasibility and technical reliability.

The performance indicators were carefully selected from a pool of advanced performance indicators [60] well known for their efficiency and ability to access, analyze and revolutionize the renewable energy space. For the benefit of this case study, we try to regenerate and adapt them as well as group them into our targets, i.e., technical, economic and environmental.

2.3.1. Technical Targets

To ensure a sustainable and reliable energy system, the technical targets primarily aim to reduce the reliance on biomass and minimize electricity imports. For this purpose, we have selected three key indicators, namely Relocation Coefficient (RC), Flexibility Factor or System Flexibility (FF) and finally, Biomass System Efficiency (BSE).

• Relocation Coefficient (RC) is defined as the measure of comparison between the ability of different technologies in the supply system flexibility. It is the ratio between the net electricity exchange between the plant and system and the electricity demand minus intermittent electricity production. This indicator essentially helps evaluate how well the energy system can adjust to changes in energy demand and production, particularly when integrating renewable energy sources that may have fluctuating output. Its formula is reported as follows:

  $$\mathrm{RC}={\displaystyle \frac{\mathrm{Net}\mathrm{electricity}\mathrm{exchange}\mathrm{between}\mathrm{plant}\mathrm{and}\mathrm{system}}{\mathrm{Electricity}\mathrm{demand}-\mathrm{intermittent}\mathrm{electricity}\mathrm{production}}}$$

  (1)
• Flexibility Factor or System Flexibility (FF) is an indicator first described by Paul Denholm and Robert M. Margolis to be the lowest hourly value over the year divided by the maximum hourly value with regard to the output of a simulation [63]. Thus, this indicator was used to assess the flexibility of the system over the year used in the simulation. We gave it a range between 0 and 1 with a value close to 0, which means the system is not flexible, and a value close to 1 means the system is flexible. In general terms, this metric helps determine how well the energy system can maintain consistent performance despite fluctuations in energy production and demand throughout the year.

  $$\mathrm{FF}={\displaystyle \frac{\mathrm{Lowest}\mathrm{hourly}\mathrm{value}\mathrm{of}\mathrm{the}\mathrm{year}}{\mathrm{Maximum}\mathrm{hourly}\mathrm{value}\mathrm{of}\mathrm{the}\mathrm{year}}}$$

  (2)
• Biomass System Efficiency (BSE) is used to assess the importance of biomass in the energy system without the transportation system [64]. This indicator was helpful in this work since it could help in the quantification and reduction of biomass in the system. To attain this, the output from synthetic fuel is subtracted from the production of all the fuel by biomass, which is then divided by the biomass used for transportation subtracted from the input amount of biomass. Essentially, this efficiency metric shows how effectively the system uses biomass resources, helping to minimize waste and maximize energy output from the available biomass.

  $$\mathrm{BSE}={\displaystyle \frac{\mathrm{Output}\mathrm{of}\mathrm{all}\mathrm{fuel}\mathrm{by}\mathrm{Biomass}-\mathrm{Output}\mathrm{from}\mathrm{synthetic}\mathrm{fuel}}{\mathrm{Input}\mathrm{Biomass}\mathrm{amount}-\mathrm{Biomass}\mathrm{used}\mathrm{for}\mathrm{transportation}}}$$

  (3)

2.3.2. Economic Targets

Economic targets are characterized by the minimization of the annual cost of the system and consist of Mismatch Compensation Factor (MCF) and Marginal Economic Efficiency (MEE).

• Mismatch Compensation Factor (MCF) was developed by Lund et al. [65] with respect to zero-energy buildings. It relates cost balance (i.e., the installed capacity of renewable energy sources where the import costs and export incomes are balanced) to energy balance (i.e., the installed capacity of renewable energy source (RES) balancing aggregated annual imports to exports from the energy system). This indicator helps measure how well the energy system can balance its energy production with its costs, ensuring that it produces enough renewable energy to meet its own needs while minimizing external energy purchases.

  $$\mathrm{MCF}={\displaystyle \frac{\mathrm{Cos}\mathrm{t}\mathrm{Balance}}{\mathrm{Energy}\mathrm{Balance}}}={\displaystyle \frac{\mathrm{Total}\mathrm{RES}\mathrm{Production}-\mathrm{RES}\mathrm{Used}}{\mathrm{Total}\mathrm{Energy}\mathrm{Demand}-\mathrm{RES}\mathrm{Used}}}$$

  (4)
• Marginal Economic Efficiency (MEE) shows how the added cost of RES contributes to the total cost of the system. It is expressed by dividing the change in the total system cost by the change in the cost of RES [60]. In simpler terms, this indicator helps assess how cost-effective the system is when adding renewable energy sources, showing whether the investment in renewable technologies leads to efficient use of resources and overall cost savings.

  $$\mathrm{MEE}={\displaystyle \frac{\Delta \mathrm{Total}\mathrm{System}\mathrm{Cost}}{\Delta \mathrm{Renewable}\mathrm{Energy}\mathrm{Sources}\mathrm{Costs}}}$$

  (5)

2.3.3. Environmental Targets

The environmental target goal is to minimize the CO₂ emissions of the system. It is also related to the system’s renewable energy sources. It has three indicators, namely, Curtailment Fraction, Marginal Primary Energy Supply and Marginal Export.

• When a system is not able to hold excess production of RES within a given period, the percentage of the RES production lost by the technology is called Curtailment Fraction (CF). When the percentage is equal to $100\%$, we say the system has the capacity to integrate the excess RES produced and vice versa. It is calculated by subtracting the realized RES production from the potential RES production, and the results are divided by the potential RES production. This indicator practically measures how much renewable energy is wasted because the system cannot fully utilize or store it, with higher curtailment indicating greater energy loss.

  $$\mathrm{CF}={\displaystyle \frac{\mathrm{Potential}\mathrm{RES}\mathrm{production}-\mathrm{Realized}\mathrm{RES}\mathrm{production}}{\mathrm{Potential}\mathrm{RES}\mathrm{production}}}$$

  (6)
• Marginal Primary Energy Supply (MPES) compares the different RES where the factors may be determined by marginal effects. Specifically, the MPES indicates how the marginal Primary Energy Supply (PES) of the system is affected by a marginal change in the PES from RES. If it is less than 1, the system cannot fully integrate marginal RES production [60]. In other therms, this indicator shows how efficiently the system can incorporate small increases in renewable energy supply, helping to assess the system’s ability to handle additional renewable energy without performance losses. This is represented by the formula below.

  $$\mathrm{MPES}={\displaystyle \frac{\Delta \mathrm{Primary}\mathrm{Energy}\mathrm{Sources}}{(\Delta \mathrm{Primary}\mathrm{Energy}\mathrm{Source}\times \mathrm{Renewable}\mathrm{Energy}\mathrm{Sources})}}$$

  (7)
• Marginal Export (ME) is used to determine the relationship between marginal export and marginal changes in PES, which are biomass-based [64].

  $$\mathrm{ME}={\displaystyle \frac{\Delta \mathrm{Export}}{\Delta \mathrm{Primary}\mathrm{Energy}\mathrm{Sources}}}$$

  (8)

These indicators were placed in a range between 0 and 1, where 0 indicates a poor performance and 1 optimal performance. Furthermore, after obtaining the results for each indicator across their respective scenarios, the data were normalized between 0 and 1 to reflect the aforementioned range. Therefore, values close to 1 represent favourable indicators, while values close to 0 represent unfavourable indicators.

Table 1. Values of the performance indicators in different scenarios, the first 10 selected as examples.

Before Normalization
Scenarios	RC	FF	CF	MCF	MPES	MEE	ME	BSE
Scenario 1	0.981722	0.025282	0.104913	0.792006	2.189719	17.02612	0.963444	0.196561
Scenario 2	0.914448	0.010581	0.049434	0.871856	2.395090	1.176235	1.064378	1.098251
Scenario 3	0.957826	0.020870	0.092879	0.917171	2.333604	1.151019	0.915653	1.099818
Scenario 4	0.967399	0.018738	0.086294	0.871856	2.406525	1.152334	0.934797	1.049301
Scenario 5	0.894655	0.009454	0.045639	0830840	2.470161	1.210174	1.081707	1.048504
Scenario 6	0.977985	0.015756	0.075988	0.830840	2.506582	1.408031	0.955969	1.043592
Scenario 7	0.960285	0.013243	0.065025	0.793491	2.537576	1.840971	1.027957	0.947787
Scenario 8	0.983609	0.017858	0.086074	0.793491	2.518343	2.206353	0.967219	0.883702
Scenario 9	0.964461	0.021068	0.100287	0.793491	2.505681	2.551852	0.928922	0.840727
Scenario 10	0.952290	0.023275	0.109848	0.793491	2.496894	2.835778	0.904580	0.812299
After Normalization
Scenarios	RC	FF	CF	MCF	MPES	MEE	ME	BSE
Scenario 1	0.850946	0.277390	0.361767	0.451797	0.111556	1.000000	0.309250	0.000000
Scenario 2	0.286335	0.151160	0.218751	0.567825	0.202887	0.056067	0.839516	0.998265
Scenario 3	0.650396	0.239502	0.330747	0.633672	0.175544	0.054565	0.058171	1.000000
Scenario 4	0.730733	0.221197	0.313773	0.567825	0.207973	0.054643	0.158749	0.944072
Scenario 5	0.120217	0.141479	0.208970	0.508226	0.236273	0.058088	0.930552	0.943190
Scenario 6	0.819577	0.195596	0.287204	0.508226	0.252470	0.069871	0.269977	0.937751
Scenario 7	0.671030	0.174015	0.258943	0.453955	0.266253	0.095655	0.648173	0.831685
Scenario 8	0.866785	0.213642	0.313203	0.453955	0.257700	0.117415	0.329079	0.760737
Scenario 9	0.706079	0.241208	0.349842	0.453955	0.252069	0.137991	0.127884	0.713158
Scenario 10	0.603932	0.260161	0.374491	0.453955	0.248161	0.154900	0.000000	0.681686

represents the first 10 indicators before normalization and after normalization to throw more light on the above statement.

2.4. TOPSIS-Based MCDA Approach

The proposed approach consists in employing the TOPSIS to evaluate and rank different scenarios based on relevant criteria. This method is herein suggested because it considers both the best and worst possible scenarios, providing a clear and balanced comparison of alternatives. The step-by-step procedure to implement the TOPSIS method is outlined as follows.

• Construct the assessment matrix: first, compile the quantitative evaluations $g_{ij}$ for each alternative i across each criterion j. This matrix provides a comprehensive overview of how each alternative performs under each criterion.
• Compute the normalized matrix, with the generic element $z_{ij}$ representing the normalized evaluation of alternative i under criterion j as:
  Normalize the matrix: next, standardize the values in the assessment matrix to make them comparable across criteria. The normalized value $z_{ij}$ for each alternative i and criterion j is calculated as:

  $$z_{ij}={\displaystyle \frac{g_{ij}}{\sqrt{{\sum }_{i=1}^n{g}_{ij}^2}}}.$$

  (9)
  This step removes the units of measurement and scales the data, ensuring that each criterion contributes equally to the analysis.
• Calculate the weighted normalized matrix: Apply the assigned weights to the normalized values to reflect the importance of each criterion. The weighted normalized value $u_{ij}$ is given by:

  $$u_{ij}=w_j\times z_{ij},\forall i,\forall j;$$

  (10)

  where $w_j$ is the weight assigned to criterion j. This step adjusts the normalized values according to the significance of each criterion.
• Determine the ideal solutions: identify the best possible (positive ideal) and worst possible (negative ideal) values for each criterion. The positive ideal solution $A^{*}$ and the negative ideal solution $A^{-}$ are defined as:

  $$A^{*}=(u_1^{*},\dots ,u_k^{*})=\{\left(\underset{i}{max}{u}_{ij}\right|j\in I^{\prime }),\left(\underset{i}{min}{u}_{ij}\right|j\in I^{″})\};$$

  (11)

  $$A^{-}=(u_1^{-},\dots ,u_k^{-})=\{\left(\underset{i}{min}{u}_{ij}\right|j\in I^{\prime }),\left(\underset{i}{max}{u}_{ij}\right|j\in I^{″})\};$$

  (12)

  where $I^{\prime }$ includes criteria to be maximized and $I^{″}$ includes criteria to be minimized. These ideal solutions serve as reference points for comparison.
• Calculate the distances to the ideal solutions: measure the distances of each alternative from the positive and negative ideal solutions. The distances $S^{*}$ and $S^{-}$ for each alternative i are computed as:

  $$S^{*}=\sqrt{\sum _{j=1}^k{(u_{ij}-u_{ij}^{*})}^2}, i=1,\dots ,n;$$

  (13)

  $$S^{-}=\sqrt{\sum _{j=1}^k{(u_{ij}-u_{ij}^{-})}^2}, i=1,\dots ,n.$$

  (14)
  These distances quantify how far each alternative is from the ideal solutions.
• Calculate the closeness coefficient: determine the closeness coefficient $C_i^{*}$ for each alternative i, which indicates its relative proximity to the ideal solutions. The closeness coefficient is calculated as:

  $$C_i^{*}={\displaystyle \frac{S^{-}}{S^{-}+S^{*}}}, 0<C_i^{*}<1,\forall i.$$

  (15)
  This coefficient shows how closely an alternative aligns with the best possible scenario while avoiding the worst.
• Rank the alternatives: finally, rank the alternatives based on their closeness coefficients in descending order. For example, in comparison between two generic alternatives i and z, if $C_i^{*}\ge C_z^{*}$, then alternative i is preferred over alternative z. This ranking helps in making informed decisions by highlighting the most favorable options.

By following these steps, the TOPSIS method provides a systematic and objective way to evaluate and rank multiple alternatives based on a set of criteria, ensuring balanced and well-informed decision-making.

After obtaining the final ranking by initially assigning equal weights to all criteria, we will conduct a sensitivity analysis by varying the weights assigned to each criterion. This analysis will help us understand the robustness of the rankings and the impact of different criteria on the overall assessment. The criteria will be grouped based on their nature, such as environmental, technical, and economic targets, allowing us to systematically explore how changes within these groups affect the final rankings.

3. Results and Discussion

This section presents the findings of the proposed MCDA across various scenarios. Firstly, the performance indicators are analyzed and grouped according to different targets, i.e., environmental, technical, and economic. The relationship between each indicator and the capacity of the key technologies is then illustrated. Subsequently, the MCDA of this section dives deep into the performance of the scenarios across all targets and the sensitivity of each scenario is analyzed. The analysed scenarios are the 52 near-optimal energy system configurations resulting from the grid search in [53].

3.1. Performance Indicators of the Energy Scenarios

This particular session of the paper is dedicated to the performance indicator results, where each scenario performance with each indicator is accessed. The indicator analysis presented in Figure 2 focuses on three primary targets: Economic, Technical, and Environmental. Each indicator within these targets offers insight into various aspects of the scenarios’ performance, allowing for a comprehensive evaluation of its efficiency, reliability, and adaptability across all the indicators.

Economic targets are depicted through the Mismatch Compensation Factor (MCF) and the Marginal Economic Efficiency (MEE) in Figure 2a. The MCF values exhibit a variation ranging from 0.8 to 1.4, indicating shifts in the cost and energy balance within the system. MCF values suggest effective compensation for mismatches in the various scenarios, indicating the best solution that is closest to 1. On the other hand, the MEE shows high variability with significant peaks over 4 in a few cases. This indicates that the share of renewable costs on overall cost varies among the different scenarios.

Technical targets are evaluated in Figure 2a using the Reliability Coefficient (RC), Flexibility Factor (FF), and Biomass System Efficiency (BSE). The RC values fluctuate moderately between 0.95 and 1.02, indicating a high-reliability level of all the scenarios with minor variability among them. This consistency underscores the system’s dependable performance across different scenarios. The FF shows values between 0 and 0.2, indicating the overall scarce flexibility of all the scenarios due to the specific characteristics of the analysed urban case study. The BSE values range from 0.8 to 1.2, suggesting that the biomass system efficiency consistently varies across scenarios. Despite this variability, values close to 1 indicate overall effectiveness in biomass utilization.

Environmental targets are then examined by means of through the Curtailment Fraction (CF), Marginal Primary Energy Supply (MPES), and Marginal Export (ME) reported in Figure 2c. The CF is characterized by values ranging between 0 and 0.5, with an important variation among the different scenarios. The highest values indicate scenarios with a moderate capacity to integrate RES. The MPES values range around 2–2.5 with a spike of 4. This consistency reflects the system’s stability in integrating renewable energy sources, with the spike indicating a potential anomaly or specific scenario causing a significant reduction. The ME values are quite stable, fluctuating slightly around 0.9 to 1.1. This stability suggests that the system maintains efficient performance in terms of energy export across different scenarios, with only minor expected fluctuations. In summary, the analyses of these indicators across economic, technical, and environmental targets reveal a complex interplay of factors influencing its performance for each scenario. The varying trends and fluctuations observed in each indicator highlight the complex dynamic of energy system configurations, emphasizing the nature of the analyzed case study.

The dependency between the eight indicators and the capacities of three key technologies (i.e., Combined Heat and Power (CHP), Heat Pump (HP), and Photovoltaic (PV)) is investigated. The relationships are illustrated through scatter plots, providing a comprehensive representation of how each indicator is affected by the installed capacities.

The analysis of the Mismatch Compensation Factor (MCF) in relation to capacities (Figure 3a) shows that the data points for CHP and HP are clustered without any pattern in relation to MCF. Instead, PV capacities, while more broadly distributed, exhibit a negative correlation with MCF. Thus, the MCF indicator, which focuses on cost balance and energy balance, appears to be relatively independent of the capacities of CHP and HP, but not from PV, suggesting that changes in PV capacity can significantly affect the marginal capacity factor.

The Marginal Economic Efficiency (MEE) indicator (Figure 3b) shows different behaviour with the different technologies. Specifically, PV and HP present a slight positive correlation with MEE, while CHP seems to not have any influence. This means that both HP and PV capacity play a positive effects on MEE.

The Reliability Coefficient (RC) indicator (Figure 3c) is again analyzed against the capacities of CHP, HP, and PV. The scatter plot indicates that all three technologies do not have a clear pattern on RC. This means that the contained variation of reliability does not depend on a single technology but on the combination of all of them together with other minor system configurations.

The Flexibility Factor or System Flexibility (FF) is also analyzed (Figure 3d). In this case, the scatter plot demonstrates a pattern for all the technologies. Specifically, CHP and HP are positively correlated with FF, with CHP showing a stronger relationship. Instead, PV is negatively correlated with FF, highlighting a clearly different impact compared with CHP and HP.

The Biomass System Efficiency (BSE) is analyzed through the scatter plot in Figure 3e). This reveals that the data points for PV have no trend but are randomly spread among various capacities over BSE values. A slight negative correlation is presented between CHP and BSE. Instead, HP shows a more evident negative relationship with biomass system efficiency. This suggests that the BSE indicator is relatively independent of PV but not from CHP and HP, which the latter has a more important impact on the biomass system’s efficiency.

Next, the Curtailment Fraction (CF) is examined. The corresponding scatter plot (Figure 3f) shows that both CHP and HP capacities exhibit a positive correlation with CF capacity. On the contrary, PV presents a negative correlation, highlighting an antithetical behaviour between PV and the other two technologies.

In examining the Marginal Primary Energy Supply (MPES) (Figure 3g), the scatter plot reveals that data points for CHP and HP are tightly clustered with no apparent variation in MPES. PV capacities show a broader range, with a slight positive relationship with MPES. This indicates that the primary energy sources and renewable energy integration are not directly affected by the installed capacities of CHP, HP, and PV, suggesting that marginal production efficiency is quite independent of the single technology capacities.

Finally, for the Marginal Export (ME) indicator (Figure 3h), the scatter plot shows that data points for CHP, HP and HP capacities have no pattern with ME.

Briefly, the dependency analysis between these performance indicators and the capacities of CHP, HP, and PV systems reveals the complexity of energy system behaviour across different technical, economic, and environmental dimensions. This finding suggests that the performance captured by these indicators is essential for accurately describing the complexity of various scenario configurations. Moreover, it highlights the importance of thoroughly analysing these solutions using an MCDA methodology.

3.2. Multi-Criteria Decision Analysis (MCDA) and Sensitivity Analysis

In this study, a MCDA was conducted to evaluate the performance of various energy scenarios for urban energy system planning. The MCDA approach employed in this analysis integrates economic, technical, and environmental targets, enabling a comprehensive assessment of each scenario’s performance under different prioritization schemes. Results are illustrated in the series of bar graphs and line charts provided, which depict the top 10 performing scenarios and the overall performance across all scenarios. In these figures, $CC$ values have been rescaled to a 0–100 range for improved interpretability.

The baseline scenario assigns equal weight ($1/3$) to each of the three macro groups of criteria, which are economic, technical, and environmental. Various indicators were distributed among the groups. The performance of this scenario is illustrated in the top 10 CC values bar graph in Figure 4a. Scenario 41 emerged as the highest-ranking scenario, consistently demonstrating superior performance across multiple criteria. This scenario’s robust performance can be attributed to its balanced approach, which ensures that none of the criteria groups are disproportionately prioritized, leading to a well-rounded energy system design.

In the environmental target scenario, 60% of the weight was allocated to environmental indicators, with the remaining 40% equally divided between technical and economic indicators. The results (Figure 4b) indicate that this scenario favours configurations that maximize environmental benefits, such as reduced emissions and efficient resource utilization. Scenario 17, which stands out, is characterized by a substantial emphasis on curtailment reduction and optimal biomass system efficiency, which are critical for minimizing the environmental footprint of the energy system.

For the technical target scenario, 60% of the weight was assigned to technical indicators, with the remaining weight equally split between economic and environmental indicators. The top-performing scenarios (Figure 4c) in this case highlight the importance of system reliability and flexibility. Scenario 41 in this case appears as the 2nd best while 32, which ranks highest in this target, reflects its superior adaptability and technical robustness, which are crucial for maintaining system stability and efficiency under varying operational conditions.

In the economic target scenario, 60% of the weight was allocated to economic indicators, with the technical and environmental indicators each receiving 20%. The top 10 scenarios in this case (Figure 4d) underline the significance of cost-effectiveness and economic efficiency in the system design. Scenario 45, which ranks highest in this scenario, showcases a highly cost-efficient configuration, effectively balancing the trade-offs between investment costs and system performance. In this case, we also see Scenario 41 performing well by placing 5th.

The comparative analysis of the different scenarios (Figure 5) shows significant variability in Closeness Coefficient (CC) values across all scenarios, indicating that no single scenario consistently outperforms the others under all target conditions. The sensitivity analysis further reveals how changes in the weighting of criteria affect the rankings of the scenarios, providing insights into the robustness of the decision-making process. Notably, Scenario 41 frequently appears in the top ranks across multiple targets, suggesting that it strikes an effective balance across all three target; technical, economic, and environmental. Unlike other scenarios that may perform well in one area but fall short in others, Scenario 41 consistently delivers high performance across multiple indicators. Its configuration offers optimal flexibility and reliability while maintaining economic efficiency and minimizing environmental impact, making it a well-rounded choice for urban energy planning. This scenario’s balanced approach aligns well with the overarching goals of transitioning to a zero-carbon future without compromising on technical stability or affordability.

This MCDA result has highlighted the importance of considering multiple perspectives when planning urban energy systems. By adjusting the weights assigned to different criteria, decision-makers can tailor the energy system design to prioritize specific objectives, whether environmental sustainability, technical reliability, or economic feasibility. The flexibility of the MCDA approach ensures that the chosen scenario aligns with broader strategic goals while accommodating the unique needs of the urban energy system under study.

Upon analyzing the sensitivity of each indicator to changes in criteria weights, we further synthesized results in

Table 2. Performance of Best Scenarios Across Different Targets.

BS	CC	ENV 60%	CC	TEC 60%	CC	ECO 60%	CC
Scenario 41	0.5994	Scenario 17	0.4868	Scenario 32	0.7190	Scenario 45	0.7729
Scenario 45	0.5936	Scenario 1	0.4777	Scenario 41	0.6992	Scenario 38	0.7389
Scenario 32	0.5790	Scenario 45	0.4580	Scenario 44	0.6903	Scenario 33	0.7197
Scenario 38	0.5762	Scenario 44	0.4537	Scenario 22	0.6802	Scenario 24	0.7098
Scenario 23	0.5648	Scenario 41	0.4536	Scenario 39	0.6783	Scenario 41	0.6993
Scenario 39	0.5596	Scenario 46	0.4385	Scenario 31	0.6730	Scenario 34	0.6847
Scenario 33	0.5517	Scenario 38	0.4369	Scenario 12	0.6715	Scenario 47	0.6781
Scenario 44	0.5502	Scenario 32	0.4313	Scenario 46	0.6708	Scenario 36	0.6613
Scenario 12	0.5485	Scenario 39	0.4300	Scenario 23	0.6657	Scenario 14	0.6601
Scenario 22	0.5476	Scenario 31	0.4277	Scenario 30	0.6564	Scenario 26	0.6552

, including the best 10 scenarios across all targets, and

Table 3. Performance of worst Scenarios Across Different Targets.

BS	CC	ENV 60%	CC	TEC 60%	CC	ECO 60%	CC
Scenario 8	0.4238	Scenario 8	0.3081	Scenario 2	0.4868	Scenario 50	0.3952
Scenario 15	0.4233	Scenario 6	0.3016	Scenario 17	0.4847	Scenario 51	0.3948
Scenario 48	0.4223	Scenario 9	0.2995	Scenario 26	0.4683	Scenario 27	0.3945
Scenario 50	0.4215	Scenario 10	0.2984	Scenario 36	0.4627	Scenario 3	0.3932
Scenario 4	0.4172	Scenario 48	0.2981	Scenario 10	0.4624	Scenario 15	0.3929
Scenario 3	0.4154	Scenario 52	0.2962	Scenario 5	0.4506	Scenario 4	0.3922
Scenario 19	0.4121	Scenario 50	0.2959	Scenario 49	0.4407	Scenario 37	0.3914
Scenario 51	0.3972	Scenario 3	0.2941	Scenario 1	0.4226	Scenario 19	0.3768
Scenario 9	0.3766	Scenario 4	0.2935	Scenario 52	0.4064	Scenario 9	0.3295
Scenario 10	0.3516	Scenario 51	0.2823	Scenario 20	0.4010	Scenario 10	0.3003

, related to the worst 10 scenarios across all targets. It is worth noting that other scenarios may be preferable if greater importance is placed on specific targets, such as Scenario 17, which performs particularly well under environmental considerations. Additionally, certain scenarios may be more suitable depending on specific constraints, such as the maximum installable capacity of a technology or budget limitations, highlighting the adaptability of the proposed methodology to different planning contexts.

4. Conclusions

In the face of rapidly advancing urbanization and the escalating demand for sustainable energy solutions, this study has significantly contributed to urban energy system planning. By introducing a novel Multi-Criteria Decision Analysis (MCDA) framework tailored to Bozen-Bolzano, we have established an integrated approach to evaluating and selecting energy scenarios.

The proposed framework balances economic, environmental, and technical targets in decision-making. The economic aspect minimizes costs while maximizing energy output, the environmental side reduces emissions and protects natural resources, and technical indicators ensure system flexibility and reliability amid fluctuating renewable energy sources. The MCDA framework allows for a careful comparison of energy options, considering the efficiency, cost, and environmental impact of different technologies, ensuring that the chosen solutions are practical and sustainable.Moreover, the tool’s flexibility makes it highly useful for decision-makers and energy policy-makers, allowing them to adjust the weighting of different criteria based on their specific priorities and constraints. However, it is important to note that this work does not delve into the integration of decision-making processes or explore the implications of energy policy, as these topics are beyond the scope of the current study.

The integration of the TOPSIS method into the MCDA framework was crucial, providing a quantifiable way to rank energy scenarios. TOPSIS simplified the decision-making process by identifying scenarios closest to the ideal solution, offering flexibility to adjust criteria weights based on evolving priorities and conditions. Scenario 41 stood out, performing well across multiple criteria, making it a compelling choice for urban planners aiming to balance economic viability, environmental sustainability, and technical reliability.

Our findings have far-reaching implications for urban areas striving to transition to $100\%$ renewable energy. The study emphasizes the necessity of a holistic approach, ensuring that selected energy systems are sustainable, economically viable, and technically sound. This is critical to achieving broader goals of climate neutrality and energy independence. Scenario analysis also played a key role in our study, allowing urban planners to anticipate uncertainties and ensure that chosen strategies are resilient and adaptable.

Beyond Bozen-Bolzano, the MCDA framework we developed can be adapted to other urban contexts, making it a versatile tool for cities facing similar challenges in energy system planning. This study contributes methodologically by demonstrating the effectiveness of integrating TOPSIS in MCDA to evaluate energy scenarios. This combination offers structured decision-making guidance while allowing for adjustments based on specific city needs, enhancing its practical application.

While our research offers valuable insights, it opens avenues for future exploration. Future research could integrate additional factors like social acceptance and technological innovation to provide a more comprehensive evaluation. Applying the MCDA framework to other cities with different socio-economic and climatic conditions could yield further insights, enhancing its generalizability.

Additionally, the dynamic nature of urban energy systems presents an opportunity to develop adaptive frameworks using real-time data and advanced computational methods like machine learning. This would allow planners to make informed, timely decisions, enhancing resilience and sustainability. Although this framework provides a comprehensive tool for evaluating energy systems, its reliance on predefined performance indicators may not fully capture future uncertainties. The study’s focus on Bozen-Bolzano, a relatively small city, also means that implementing the proposed methodology on larger and complex cities can bring challenging issues, such as data collection and energy systems modelling.

Future updates should address these limitations by incorporating real-time data and diverse socio-political factors. In conclusion, this study represents a significant advancement in urban energy planning. The framework offers a novel, adaptable tool to support the transition to $100\%$ renewable energy systems, with insights that are relevant not only to Bozen-Bolzano but to urban energy planning worldwide.