An integrated multi-objective optimization and multi-criteria decision-making model for optimal planning of workplace charging stations*

This study addresses the optimal planning of electric vehicle charging infrastructure at workplaces. As the optimal planning for a given workplace can involve various criteria that comprise conflicting single objectives, this study proposes a new integrated multi-objective optimization and multi-criteria decision-making (MCDM) model for determining the most suitable electric vehicle supply equipment (EVSE) configuration. This approach combines the advantage of multi-objective optimization, which yields Pareto solutions, with an improved MCDM model. The latter is used to evaluate the Pareto frontier to find the best performing solution by enabling the station owners to use linguistic variables for weighting the decision-making variables. The conventional weighted aggregated sum product assessment (WASPAS) method is improved by introducing the Dombi Bonferroni functions in the proposed model making it more flexible as compared to its counterparts. In the final step, the selected solutions are ranked by reapplying the MCDM model. A case study is performed based on collected charging data from a workplace. To validate the proposed model, a comparison against four alternative MCDM models is performed. It is demonstrated that the proposed model yields very close ranking order as the alternative approaches. Among five EVSE options, DC fast charging is found to be the best while AC Level-2 EVSE (19.2/22 kW) is found to be the least attractive option. Sensitivity analysis shows the robustness of the ranking results in response to changing weightings of the model coefficients.

Automated summary

This study introduces a new integrated multi-objective optimization and multi-criteria decision-making (MCDM) model to determine the most suitable electric vehicle supply equipment (EVSE) configuration at workplaces. By combining the advantages of multi-objective optimization and an improved MCDM model, the proposed approach allows station owners to use linguistic variables for weighting decision-making variables, resulting in finding the best performing solution.