Accelerating Two-Phase Multiobjective Evolutionary Algorithm for Electric Location-Routing Problems

Chao Wang,Rongrong Liu,Yansen Su,Xingyi Zhang

DOI: https://doi.org/10.1109/tits.2023.3266807

IF: 8.5

2023-01-01

IEEE Transactions on Intelligent Transportation Systems

Abstract:Electric location-routing problem is a challenging problem consisting of the optimization of electric vehicle routing and charging facility location, simultaneously. Existing algorithms generally adopt the two-phase search strategy to alternately optimize the routing and the location. However, they are usually criticized for the inefficiency as the problem scale increases. In order to improve the search efficiency in each phase, we propose an accelerating two-phase multiobjective evolutionary algorithm, where the learning method is used to mine the useful information from the historical search process to generate the high-quality routing and location offspring. To be specific, in the routing optimization phase, an interpolation method is developed to extract the frequent visiting orders existing in the historical best solutions. These frequent visiting orders are used to create potential routing offspring that can accelerate the convergence toward the optimal solutions. In the location optimization phase, a surrogate model is used to approximatively represent the relationship from routing to location, which can directly output a promising location scheme for a given routing offspring and thus reduce the optimization time. Experimental results on different scales of test instances demonstrate the competitiveness of the proposed algorithm in comparison with several state of-the-art algorithms, including four widely used heuristic algorithms and two multiobjective evolutionary algorithms.

engineering, electrical & electronic,transportation science & technology, civil