Funded by: EPSRC
This project will develop a proof-of-concept planning model for central planners to optimally locate electric vehicles (EVs) charging infrastructure under the risk of disruption to charging points (i.e. unexpected failure, technical faults or breakdowns). The aim of the model will be to maximise total expected traffic volume of EVs that can be charged by an unreliable integrated charging network, via both both static and dynamic systems.
A robust mixed-integer non-linear programming (MINLP) model for this problem will be formulated. Queuing theory equations will be incorporated into the model to account for the stochastic nature of demand both spatially and over time (e.g. peak versus off-peak periods). The model will be further generalised to a multi-period planning problem given limited periodic budgets. The model will be linearised so that it can be solved using a general-purpose solver. Finally, an efficient metaheuristic algorithm will be developed to solve the large-scale real-world instances within a reasonable computational time.
A case study of the road network in the UK will be used to assess the accuracy and performance of the linearised optimisation model and the metaheuristic algorithm. Other outputs will be the creation of test datasets and journal articles. Codes of the model and algorithm, and test datasets will also be made available to the community of Operational Research so that other researchers and practitioners (e.g., National Grid) can use them in their own case studies.
Professor Jesse O’Hanley, from the CeLSA at Kent business school is the project CO-I.