An EDA-based method for solving electric vehicle charging scheduling problem under limited power and maximum imbalance constraints. (October 2022)

• Record Type: Journal Article
• Authors: Shahmoradi, Hadi; Esmaelian, Majid; Karshenas, Hossein
• Is Part Of: Computers & industrial engineering. Volume 172:Part A(2022)
• Publication Date: 2022-10
• Subjects: Electric Vehicle Charging Scheduling Problem (EVCSP) -- Balance constraint -- Estimation Distribution Algorithm (EDA) -- Markov Network-based EDA -- Mallows Model-based EDA -- Constraint Programming (CP); Engineering -- Data processing -- Periodicals; Industrial engineering -- Periodicals; 620.00285
• DOI: 10.1016/j.cie.2022.108544
• Languages: English
• ISSNs: 0360-8352

Abstract

Graphical abstract: Highlights: The variable of vehicle assignment to charging lines, makes scheduling more realistic. Estimation of distribution algorithm can handle relations in charging scheduling. A hybrid Markov network and Mallows model is proper for solving charging scheduling. A Constraint programming approach provides optimal solution for charging scheduling. Abstract: Electronic vehicles (EVs) are receiving increasing attention to addressing global warming challenges since fossil fuel is replaced with fuel cell technology. Hence, new challenges arise as demands have increased for using EVs. One of these challenges is the long waiting time of charging EVs spent in queues, especially during peak hours. So, in this study, we aim to propose an efficient method for the electric vehicle charging scheduling problem (EVCSP), which an actual charging station inspires. The most important constraint in this problem is balancing power consumption between charging lines, leading to a limited number of devices that can be charged simultaneously. Also, in this problem, EVs may have interrelationships with each other during the scheduling procedure. So, the estimation of distribution algorithm (EDA) as a competent method in handling the possible relations among decision variables is applied in our proposed hybrid EDA-based solving method. Our proposed method comprises two EDAs, a Markov network-based EDA and a Mallows model-based EDA. It achieves an appropriate schedule and charging line assignment simultaneously while minimizing the total tardiness considering problem constraints. We compared our method with a constraint programming (CP) model and the state-of-art meta-heuristic methods in terms of the objective function value by simulation on a benchmark dataset. Results from the experimental study show significant improvement in solving the introduced EVCSPs.