A Multi-Dimensional Sustainability Framework for Dynamic Ridesharing Platforms: Integrating Hybrid Optimization to Address Environmental, Social, and Economic Impacts

A Multi-Dimensional Sustainability Framework for Dynamic Ridesharing Platforms: Integrating Hybrid Optimization to Address Environmental, Social, and Economic Impacts

¹Ribadu Rukkaiyatu Bashir; ²Odekunle Remilekun Mathew; ³Momoh Abdulfatai Atte; ⁴Eli Joel, ⁵Binibonori Salihu Tanko

^{1,2 & 3}Modibbo Adama University, Yola Adamawa State, Nigeria

⁴Post primary school Management Board Yola, Adamawa State Nigeria, Nigeria

⁵Federal Polytechnic, Mubi Adamawa State, Nigeria

DOI: https://dx.doi.org/10.47772/IJRISS.2025.908000527

Received: 14 August 2025; Accepted: 21 August 2025; Published: 19 September 2025

ABSTRACT

Urban mobility systems in emerging cities are increasingly strained by rising greenhouse gas emissions, inefficient vehicle utilization, and unsustainable travel demand patterns. Dynamic ridesharing, when guided by intelligent optimization, offers significant potential to reduce vehicle miles travelled (VMT), lower energy consumption, and curb urban emissions. This study introduces an environmentally focused hybrid multi-objective optimization framework that integrates evolutionary algorithms (NSGA-II and SPEA2) with machine learning models (Neural Networks and Reinforcement Learning) to optimize real-time ridesharing decisions under diverse urban demand conditions. Implemented in a MATLAB-based simulation environment, the framework captures realistic geographic and temporal variability in ride requests. Three urban demand scenarios low, medium, and high were analyzed to evaluate the system’s responsiveness and adaptability. Core optimization components were modularized for scalability, and hybridization ensured balanced trade-offs across environmental and operational objectives. Simulation results demonstrated that the hybrid model outperformed both the baseline (no ridesharing) and traditional ridesharing setups, achieving up to 35% reduction in VMT, significant cost savings, and improved travel time control. In high-demand settings, the model further reduced system-wide trip costs while preserving operational fairness. Survey responses from over 100 participants indicated high public acceptance of environmentally sustainable ridesharing systems, particularly those emphasizing equity, affordability, and low-impact routing. The research underscores the importance of embedding sustainability criteria within urban mobility algorithms and offers a computationally efficient, behaviorally informed, and scalable model suited for African cities grappling with congestion, emissions, and transport equity. The proposed framework serves as a replicable tool for urban planners and mobility service providers aiming to balance environmental sustainability, economic efficiency, and social inclusion in the design of next-generation transport systems.

Keywords: Dynamic ridesharing, sustainability, hybrid optimization, environmental impact, smart mobility, urban transport, equity, emissions reduction

INTRODUCTION

Urban mobility systems in emerging economies are undergoing rapid transformation amidst rising challenges and evolving technological opportunities. Many cities across Africa, Asia, and Latin America are experiencing explosive population growth, increasing motorization, and sprawling urban development all of which place intense pressure on transportation infrastructure and services (UN-Habitat, 2020). The consequences are visible: mounting traffic congestion, elevated levels of air and noise pollution, longer commute times, and growing disparities in transport access. Private vehicle ownership continues to rise in these contexts, often due to the inadequacy of affordable and reliable public transportation options (Pojani & Stead, 2015; World Bank, 2021). As a result, urban transport contributes significantly to greenhouse gas (GHG) emissions, energy consumption, and socio-spatial inequality.

In response to these concerns, dynamic ridesharing platforms which facilitate real-time, on-demand shared travel among passengers with similar routes have emerged as a compelling mobility solution. These systems aim to improve vehicle occupancy, reduce total vehicle kilometers traveled (VKT), and offer cost-effective alternatives to both private car ownership and traditional ride-hailing services (Shaheen & Chan, 2016; Agatz et al., 2012). When appropriately designed, ridesharing can play a crucial role in transitioning toward low-carbon, equitable, and inclusive urban mobility systems.

However, most current ridesharing systems are narrowly optimized for operational performance focusing on metrics such as travel time, trip cost, and system throughput while often neglecting critical dimensions of environmental sustainability, social equity, and economic resilience (Firnkorn & Müller, 2015; Feigon & Murphy, 2016). The unintended consequences of such narrow optimization include an increase in unbalanced demand across neighborhoods, detours that disproportionately affect disadvantaged users, and limited consideration for environmental trade-offs such as congestion-induced emissions.

Given the complexity and interdependence of sustainability goals in urban transport systems, a multi-objective and interdisciplinary approach is necessary one that considers the synergies and trade-offs between environmental, social, and economic dimensions (Litman, 2021). Urban mobility solutions must align with broader global goals, particularly the United Nations Sustainable Development Goals (SDGs) notably SDG 11 (Sustainable Cities and Communities) and SDG 13 (Climate Action).

To address these challenges, this study introduces a novel Hybrid Multi-Objective Optimization Framework (H-MOOF) that integrates evolutionary algorithms (NSGA-II and SPEA2) and machine learning models (including neural networks and reinforcement learning) to guide dynamic ridesharing decisions under diverse and evolving urban conditions. Unlike conventional algorithms, the proposed framework balances conflicting objectives such as minimizing vehicle miles traveled (VMT), trip costs, and GHG emissions, while also considering social fairness, system scalability, and economic efficiency.

Moreover, the research develops and applies a comprehensive sustainability assessment framework, capturing key performance indicators (KPIs) related to accessibility, equity, affordability, environmental impact, and platform economics. This integrated framework is designed to evaluate and guide the performance of dynamic ridesharing systems, particularly within the urban mobility landscape of African cities, where issues of transport exclusion, infrastructure constraints, and environmental vulnerability are acute (Barter, 2021; Odufuwa et al., 2020).

The insights gained from this study are intended to support evidence-based policymaking, platform design, and urban mobility planning. Ultimately, this work contributes a replicable model for deploying intelligent and sustainable ridesharing platforms that align with long-term climate and equity goals in emerging urban regions.

Statement of the Problem

Urban transport in rapidly growing cities is strained by population growth, weak public systems, rising private vehicle use, and inefficient informal networks leading to congestion, higher costs, emissions, and mobility inequities. Dynamic ridesharing offers potential solutions, but in Nigeria most systems remain informal, outdated, and poorly scalable. Conventional optimization methods also fail to balance key goals such as reducing travel time, costs, and vehicle miles.

Advanced tools like evolutionary algorithms and machine learning, though effective elsewhere, are underutilized, and real-time data is rarely leveraged. Additionally, user concerns such as safety, trust, affordability, and gender sensitivity are often neglected. To address these gaps, this study introduces a scalable hybrid optimization framework for Adamawa State that integrates real-time data, advanced computational techniques, and sustainability metrics to improve efficiency, reduce environmental impacts, and ensure equitable mobility access.

Objectives

The primary objective of this study is to:

Develop a comprehensive framework to assess the social, economic, and environmental impacts of dynamic ridesharing platforms, and conduct studies to evaluate the sustainability and equity implications of these platforms.

Significance of the Study

This study advances dynamic ridesharing by introducing a scalable and sustainability-driven optimization framework that integrates evolutionary algorithms, machine learning, and real-time mobility data. Unlike traditional models focused only on efficiency, the framework adopts a multi-objective approach that balances environmental, economic, and social goals by minimizing Vehicle Miles Traveled (VMT), Travel Time (TT), and Trip Cost (TC).

Its novelty lies in a comprehensive sustainability assessment that considers emissions, equity, affordability, safety, gender inclusion, and user satisfaction, making the model both adaptive and user-centered. The findings provide a policy-relevant foundation for designing ridesharing platforms that are technologically advanced, environmentally responsible, socially equitable, and economically viable supporting long-term sustainable mobility agendas.

Scope of the Study

The scope of this study is focused on the design, simulation, and validation of a hybrid multi-objective optimization framework specifically tailored for dynamic ridesharing systems. The study will utilize synthetic and real-world urban mobility data from selected cities that present varying transportation challenges such as demand volatility, congestion, and infrastructure constraints.

The framework will be tested under multiple demand conditions low, medium, and high and evaluated based on its performance in minimizing VMT, TT, and TC, while also assessing its broader environmental, social, and economic implications. These indicators will serve as benchmarks for sustainability performance, using both simulation results and user behavior data obtained through structured surveys.

This research is limited to the ridesharing sector and does not extend to freight logistics, public transportation planning, or comprehensive urban land-use strategies. Although it incorporates interdisciplinary methodologies drawn from operations research, computational intelligence, and behavioral modeling, its application domain remains squarely within shared urban passenger mobility.

LITERATURE REVIEW

Environmental Sustainability and Urban Transport

The transportation sector is a major contributor to global climate change, accounting for approximately 25% of global CO₂ emissions (IEA, 2021). Within this sector, urban transport is particularly responsible due to high vehicular density, inefficient traffic patterns, and dependence on fossil-fuel-powered private vehicles. As cities grow, the negative externalities of urban transport such as air pollution, noise, congestion, and carbon emissions—are becoming increasingly unsustainable, especially in developing countries where regulatory enforcement and clean transport infrastructure are often weak (Sims et al., 2014).

In this context, dynamic ridesharing platforms offer a potential solution to reduce vehicle miles traveled (VMT) and mitigate environmental impacts by encouraging multiple passengers to share trips with similar origins or destinations. By increasing vehicle occupancy rates and reducing the number of single-occupancy vehicles, ridesharing can lower per capita fuel consumption and emissions per passenger-kilometer (Shaheen et al., 2016; Li et al., 2021). Furthermore, integration of these platforms with electric vehicle (EV) fleets and public transport systems can amplify their environmental benefits (Zhang et al., 2021). However, the environmental gains from ridesharing are not automatic they depend heavily on system design, user participation, and optimization algorithms that minimize detours and idle times while maximizing efficiency.

Despite their potential, ridesharing platforms can also generate rebound effects, where increased convenience may encourage more travel or even induce demand from users who would otherwise walk, cycle, or use public transit (Cohen et al., 2018). Therefore, ridesharing must be managed within a sustainability framework that prioritizes emissions reduction as a core objective.

Multi-Objective Optimization in Ridesharing

Dynamic ridesharing systems operate in complex, real-time environments where multiple, often conflicting, objectives must be balanced. Traditional optimization approaches in the domain have primarily focused on minimizing operational metrics such as total travel time (TT), waiting time, and trip cost (TC) (Agatz et al., 2012). However, such single-objective models neglect critical dimensions like system-wide energy efficiency, equity of service allocation, and environmental sustainability.

To address these limitations, multi-objective optimization (MOO) frameworks have been proposed. These frameworks utilize evolutionary algorithms such as the Non-dominated Sorting Genetic Algorithm II (NSGA-II) (Deb et al., 2002) and the Strength Pareto Evolutionary Algorithm 2 (SPEA2) (Zitzler et al., 2001) to simultaneously optimize several objectives. These algorithms generate a Pareto front of solutions, providing decision-makers with a range of trade-offs between competing criteria like minimizing VMT, reducing CO₂ emissions, and maximizing user satisfaction.

In recent years, reinforcement learning (RL) has also been applied in ridesharing contexts to support dynamic decision-making under uncertainty. RL agents learn optimal dispatching and routing strategies through continuous feedback from the environment, making them suitable for scalable, adaptive systems that evolve with urban demand patterns (Elhenawy et al., 2020; Xu et al., 2021). Hybrid models that integrate RL with MOO algorithms can capture both the predictive intelligence of machine learning and the exploratory strength of evolutionary optimization—enabling better system-wide sustainability performance.

Despite promising results, most MOO applications in ridesharing have yet to incorporate social dimensions such as inclusiveness and accessibility for marginalized groups, pointing to a critical research gap.

Sustainability Assessment Frameworks

As the scope of urban transport challenges expands beyond efficiency to include broader sustainability goals, there is an increasing need for comprehensive assessment frameworks that evaluate transport systems across environmental, economic, and social dimensions—the so-called “triple bottom line” (Elkington, 1997).

While many studies focus on environmental impacts such as emissions and fuel consumption—there is limited integration of social and equity metrics into ridesharing evaluations. For instance, ridesharing platforms may inadvertently exclude users who lack smartphones, bank accounts, or digital literacy widening the mobility gap (Smith & Levere, 2020). Similarly, economic accessibility remains underexplored; fare structures that are unaffordable for low-income users can undermine the inclusive potential of shared mobility.

Recent efforts to address these gaps have proposed sustainability indicators tailored to urban mobility, including accessibility indices, equity of service coverage, driver income stability, and environmental justice outcomes (Litman, 2021; Shaheen & Cohen, 2020). However, few frameworks operationalize these indicators into real-time ridesharing algorithms a disconnect that limits the potential of these platforms to contribute meaningfully to SDGs such as SDG 10 (Reduced Inequalities) and SDG 13 (Climate Action).

This study seeks to bridge this gap by embedding a multi-dimensional sustainability framework into a hybrid optimization model, enabling data-driven assessment and improvement of ridesharing platforms on a systems level.

METHODOLOGY

This study employs a hybrid computational–behavioral approach to develop and evaluate a real-time, environmentally sustainable ridesharing framework. The methodology integrates advanced machine learning and evolutionary optimization techniques with a structured sustainability assessment and user perception analysis. The methodological workflow is organized into four major components: framework architecture, simulation design, sustainability and equity assessment, and validation/user feedback.

Framework Architecture

The core of the study is the development of a Hybrid Multi-Objective Optimization Framework (H-MOOF) designed to dynamically optimize ride-matching decisions while simultaneously addressing environmental, social, and economic objectives. The architecture combines three distinct computational components:

a. Evolutionary Algorithms for Multi-Objective Optimization

The framework incorporates two well-established evolutionary algorithms:

1. Non-Dominated Sorting Genetic Algorithm II (NSGA-II) (Deb et al., 2002)
2. Strength Pareto Evolutionary Algorithm 2 (SPEA2) (Zitzler et al., 2001)

These algorithms generate Pareto-optimal solutions that trade off among three key system-level objectives:

1. Vehicle Miles Traveled (VMT): a proxy for emissions and road congestion
2. Travel Time (TT): a measure of user experience and system efficiency

• Trip Cost (TC): reflecting user affordability and platform economic performance

The use of multiple algorithms increases solution diversity and robustness, with the final solution set selected based on real-time demand characteristics and performance priorities.

b. Neural Networks for Demand Forecasting

A Multi-Layer Perceptron (MLP) neural network model is integrated into the framework to perform short-term demand forecasting, based on:

• Temporal features: hour-of-day, day-of-week, holidays
• Spatial features: demand hotspots, proximity to transit nodes, and residential density

This predictive layer enables proactive resource allocation and dynamic ride clustering, improving system responsiveness and reducing idle time.

c. Reinforcement Learning for Adaptive Policy Adjustment

To optimize decisions under real-time variability, a Q-learning-based reinforcement learning (RL) agent is deployed to fine-tune ride-matching and routing strategies in high-demand or congested scenarios. The RL agent continuously interacts with the simulated environment, receiving feedback on rewards such as:

• Emissions saved per ride
• Detour penalty avoided
• Service fairness index

This adaptive component ensures the framework remains flexible and learning-driven, even under non-stationary demand conditions.

Simulation Environment

The hybrid framework was implemented in a MATLAB simulation environment designed to replicate realistic urban mobility dynamics. The simulation was structured around three representative demand scenarios:

a. Demand Scenarios

1. Low-Demand Scenario: simulating off-peak hours (e.g., late night, early morning), low ride request frequency
2. Medium-Demand Scenario: regular daytime periods with moderate ride requests

• High-Demand Scenario: peak hours with spatial clustering (e.g., commercial centers, schools) and temporal surges

Each scenario accounted for:

1. User location distribution based on a spatial probability matrix
2. Temporal variation in request arrival rates and ride durations
3. Multi-modal transport options, including private car usage, conventional rideshare, and public transit proximity

The platform continuously monitored and adjusted for ride-matching efficiency, trip scheduling, and vehicle dispatching.

Sustainability and Equity Impact Assessment

To holistically evaluate the proposed framework’s performance, a multi-dimensional sustainability and equity assessment framework was developed. The framework aligns with global sustainability goals ( SDG 11, 12, and 13) and incorporates indicators across three dimensions (See Table 1).

Table 1: Sustainability and Equity Impact Assessment

a. Environmental Metrics

Emissions were calculated using the COPERT-based emission factor model, adjusted for urban stop-and-go traffic conditions. VMT was logged in real-time, and vehicle occupancy rates were tracked for all ride segments.

b. Social Metrics

Accessibility was assessed by service coverage in low-income and peri-urban zones.

1. Affordability was gauged by the percentage of rides falling below a cost threshold of the city’s average hourly wage.
2. Gender inclusion was proxied through platform preference settings and route safety scores.

• Detour frequency was quantified as a fairness metric, particularly for riders traveling from marginalized zones.

c. Economic Metrics

1. Trip costs and driver revenues were calculated per match scenario.
2. Platform profitability was simulated using simplified commission models to assess system scalability.

Validation and User Perception Study

To validate simulation outcomes and assess societal acceptance, the study employed a two-pronged validation strategy:

a. Comparative Scenario Evaluation

Three comparative system configurations were analyzed:

1. Baseline Scenario (No ridesharing): private vehicle or one-passenger taxis
2. Conventional Ridesharing (No optimization): basic nearest-neighbor match

• Hybrid Optimized Ridesharing (H-MOOF): full integration of MOO + ML + RL

b. Public Attitude and Perception Survey

A structured survey (n=100) was administered across three urban centers to understand:

1. Awareness of ridesharing platforms
2. Preferences for sustainable mobility

• Concerns about affordability, safety, and equity

1. Willingness to use platforms offering low-emission, shared rides

The survey combined ranking tasks, and open-ended feedback, with a focus on identifying behavioral and attitudinal enablers of sustainable ridesharing adoption.

Survey insights were integrated into model refinement by adjusting fairness weights and inclusion constraints within the optimization process.

RESULTS AND FINDINGS

This section presents findings from simulation experiments and user surveys, focusing on optimization performance, environmental impact, social equity, and public perception. The Hybrid Multi-Objective Optimization Framework (H-MOOF) was evaluated against baseline (no ridesharing) and conventional (basic pooling) models to assess its effectiveness in promoting sustainable urban mobility.

Optimization Performance

The hybrid framework significantly improved Vehicle Miles Traveled (VMT), Travel Time (TT), and Trip Cost (TC). It reduced VMT by up to 35%, lowered average TT by 18% under high demand, and decreased system-wide trip costs by 25%, while driver earnings improved by 12%.

Validation compared three scenarios: baseline (private trips), conventional ridesharing, and the hybrid framework. As shown in Table 2, the baseline had the highest VMT but the lowest TT and TC since each passenger used a dedicated vehicle. Conventional ridesharing cut VMT substantially but increased TT and TC due to inefficient routing. In contrast, the hybrid framework achieved a similar VMT reduction while limiting TC increases to only 6% compared with baseline, far better than the 31% rise under conventional pooling.

Table 2 summarizes performance across the three scenarios:

The results in Table 2 highlight the trade-offs inherent in ridesharing. While conventional ridesharing reduced VMT significantly, it also increased TT and TC due to inefficient routing. The hybrid framework achieved comparable VMT reductions but limited cost increases to only 6% above baseline, demonstrating an optimized balance between efficiency and user impact.

Analysis of trade-offs and practical implications

The results reveal clear trade-offs inherent in ridesharing systems, which are effectively managed by the proposed hybrid optimisation framework. Figures 1 and 2 visually summarize these performance comparisons.

Figure 1: Performance Metric Across Scenarios.

Figure 2: Percentage Reduction Metric Across Scenarios.

Environmental Impact

The framework reduced system-wide CO₂ emissions by 33% and fuel consumption by 28%, largely due to lower VMT and increased occupancy. Average vehicle occupancy rose from 1.1 to 2.9 passengers per trip, supporting international climate goals such as SDG 11 (Sustainable Cities) and SDG 13 (Climate Action).

Social and Equity Outcomes

The hybrid model enhanced service accessibility in underserved areas and incorporated gender-sensitive routing options, leading to a 22% increase in female ridership. Intelligent route pruning reduced excessive detours, creating fairer travel time distribution among passengers and improving satisfaction. These results confirm the potential of optimized ridesharing to advance inclusivity and equity in line with SDG 10 (Reduced Inequalities).

Public Perception and User Insights

A survey of 100 urban residents revealed high readiness for ridesharing adoption. Eighty-five percent of respondents were familiar with ridesharing, and most expressed willingness to share rides, especially when cost and environmental benefits were emphasized. Around 70% accepted an additional 6–10 minutes waiting time and detours of up to 20%. Users also identified safety and trust as critical priorities, favoring features such as real-time tracking and gender-based preferences. Incentives such as discounted fares and flexible payment options were highlighted as drivers of participation. These behavioral insights reinforce the feasibility and user-centered nature of the framework.

Impact Assessment

The overall impact assessment confirmed multi-dimensional benefits of H-MOOF. Environmentally, the 35% VMT reduction significantly lowered emissions and energy use. Economically, users experienced cost savings while drivers earned more, creating a balanced ecosystem. Socially, inclusivity and equitable access were enhanced through demand-sensitive routing and gender-aware features. Collectively, the results demonstrate that H-MOOF can effectively support sustainable, efficient, and inclusive urban mobility.

SUMMARY, CONCLUSION AND RECOMMENDATION

This study developed and validated a hybrid multi-objective optimization framework that integrates evolutionary algorithms and machine learning to address urban mobility challenges. The framework minimized VMT, TT, and TC while promoting environmental, economic, and social sustainability.

Key outcomes include a 35% reduction in VMT, significant cost savings for users, and increased driver earnings. Although TT increased slightly due to pooling, the trade-off was acceptable and offset by broader efficiency and inclusivity gains. Survey results further confirmed strong user willingness to adopt optimized ridesharing when affordability, safety, and trust features are embedded.

In conclusion, the findings demonstrate that the proposed H-MOOF is computationally feasible, environmentally responsible, and socially inclusive. It offers a scalable pathway for ridesharing platforms to enhance urban transport systems in emerging cities, aligning with global sustainability and digital transformation agendas.

Recommendations

Based on the outcomes of this study, several actionable recommendations are offered for policymakers, urban planners, mobility service providers, and researchers:

1. Policymakers should embed multi-objective optimization models within public and private ridesharing systems to achieve urban sustainability targets related to emissions reduction, traffic decongestion, and transport equity.
2. Ridesharing platforms are encouraged to implement hybrid AI systems capable of responding to demand fluctuations, user preferences, and routing constraints in real time, ensuring system resilience and user satisfaction.
3. Effective user engagement, digital literacy, and trust-building strategies including incentives, transparency, and safety features are essential to increasing public willingness to participate in shared mobility services.
4. Governments and technology developers should invest in mobility data systems that allow for continuous performance monitoring, sustainability tracking, and integration of ridesharing platforms into broader smart city ecosystems.

Contribution to Knowledge

This study contributes novel insights and methodological innovations to the evolving field of intelligent urban transport and shared mobility, particularly within emerging economy contexts. Key contributions include:

1. Hybrid Optimization Model Development introduced a new algorithmic framework that integrates evolutionary algorithms (NSGA-II, SPEA2) with machine learning (Neural Networks, Reinforcement Learning) for dynamic ridesharing optimization.
2. Comprehensive Performance Validation provided rigorous validation of optimization outcomes across three scenarios baseline, conventional, and hybrid filling a gap in previous ridesharing literature that often overlooks comparative benchmarking.
3. Sustainability Impact Assessment developed and applied a multi-dimensional sustainability assessment framework encompassing environmental (e.g., CO₂ emissions, VMT), social that is equity, gender inclusion and economic indicators.
4. Behaviorally Informed Design Integration successfully linked computational models with real-world user behavior data, offering a replicable methodology for integrating optimization algorithms with public attitudes and adoption preferences in urban mobility studies.

In conclusion, this research bridges the gap between theoretical optimization and practical implementation in the field of dynamic ridesharing. It offers a scalable, equitable, and environmentally conscious blueprint for urban mobility transformation especially relevant for rapidly urbanizing regions in Nigeria and beyond.

REFERENCES

1. Agatz, N., Erera, A., Savelsbergh, M., & Wang, X. (2012). Optimization for dynamic ride-sharing: A review. European Journal of Operational Research, 223(2), 295–303. https://doi.org/10.1016/j.ejor.2012.05.028
2. Agatz, N., Erera, A., Savelsbergh, M., & Wang, X. (2012). Optimization for dynamic ride-sharing: A review. European Journal of Operational Research, 223(2), 295–303. https://doi.org/10.1016/j.ejor.2012.05.028
3. Barter, P. (2021). Informal transport and ridesharing in African cities: An opportunity for inclusive mobility? Sustainable Cities and Society, 66, 102691. https://doi.org/10.1016/j.scs.2020.102691
4. Cohen, A., & Shaheen, S. (2018). Planning for Shared Mobility. Transportation Research Board. https://doi.org/10.17226/24920
5. Deb, K., Pratap, A., Agarwal, S., & Meyarivan, T. (2002). A fast and elitist multiobjective genetic algorithm: NSGA-II. IEEE Transactions on Evolutionary Computation.
6. Elhenawy, M., Elbery, A., & Rakha, H. (2020). Reinforcement learning-based dynamic ridesharing system. Transportation Research Part C: Emerging Technologies, 120, 102802. https://doi.org/10.1016/j.trc.2020.102802
7. Elkington, J. (1997). Cannibals with Forks: The Triple Bottom Line of 21st Century Business. Capstone.
8. Feigon, S., & Murphy, C. (2016). Shared Mobility and the Transformation of Public Transit. Transit Cooperative Research Program (TCRP) Report 188. Transportation Research Board.
9. Firnkorn, J., & Müller, M. (2015). Free-floating electric carsharing–favorable potential impact on car ownership and GHG emissions in urban areas. Transportation Research Part D: Transport and Environment, 31, 272–284. https://doi.org/10.1016/j.trd.2014.07.022
10. IEA (2021). Tracking Transport 2021. International Energy Agency. https://www.iea.org/reports/tracking-transport-2021
11. Li, M., Wang, D., & Li, Y. (2021). Evaluating environmental benefits of ridesharing with electric vehicles in Beijing: A simulation-based approach. Sustainable Cities and Society, 66, 102699. https://doi.org/10.1016/j.scs.2020.102699
12. Litman, T. (2021). Evaluating Transportation Equity. Victoria Transport Policy Institute.
13. Litman, T. (2021). Evaluating Transportation Equity: Guidance for Incorporating Distributional Impacts in Transportation Planning. Victoria Transport Policy Institute.
14. Odufuwa, B. O., Akinmoladun, I. O., & Oladosu, O. O. (2020). Urban mobility, equity and sustainability in Nigeria: The need for reform. Transportation Research Procedia, 48, 2560–2574. https://doi.org/10.1016/j.trpro.2020.08.191
15. Pojani, D., & Stead, D. (2015). Sustainable urban transport in the developing world: Beyond megacities. Sustainability, 7(6), 7784–7805. https://doi.org/10.3390/su7067784
16. Shaheen, S., & Chan, N. (2016). Mobility and the Sharing Economy: Potential to facilitate the first- and last-mile public transit connections. Built Environment, 42(4), 573–588.
17. Shaheen, S., & Cohen, A. (2020). Mobility on demand (MOD) and mobility as a service (MaaS): Early understanding of shared mobility impacts and public transit partnerships. In Advances in Transport Policy and Planning (Vol. 6, pp. 47–70). Elsevier.
18. Shaheen, S., Chan, N., Bansal, A., & Cohen, A. (2016). Shared mobility: Definitions, industry developments, and early understanding. UC Berkeley: Transportation Sustainability Research Center.
19. Shaheen, S., Cohen, A., & Zohdy, I. (2016). Shared Mobility: Current Practices and Guiding Principles.S. Department of Transportation.
20. Sims, R., Schaeffer, R., Creutzig, F., et al. (2014). Transport. In: Climate Change 2014: Mitigation of Climate Change. Contribution of Working Group III to the IPCC Fifth Assessment Report.
21. Smith, A., & Levere, M. (2020). Ridesharing and Inclusive Mobility: Trends, Equity Gaps, and Policy Solutions. Urban Institute. https://www.urban.org/research/publication/ridesharing-and-inclusive-mobility
22. World Bank (2021). Transport Sector Emissions: Pathways to Decarbonization.
23. World Bank. (2021). The Hidden Wealth of Cities: Creating, Financing, and Managing Public Spaces.
24. Zhang, H., Wang, H., & Zhang, Y. (2021). Sustainable ridesharing systems: A review of algorithmic, environmental, and equity aspects. Journal of Cleaner Production, 278, 123808. https://doi.org/10.1016/j.jclepro.2020.123808
25. Zitzler, E., Laumanns, M., & Thiele, L. (2001). SPEA2: Improving the strength Pareto evolutionary algorithm. Technical Report 103, ETH Zurich.