An improved particle swarm optimization with particle refactor operator for perishable food delivery problems by electric vehicles

Abstract

ABSTRACT The increasing demand for perishable food and the popularity of electric vehicles have promoted the integration research of perishable food delivery services and electric vehicles. Aiming at minimizing the total delivery cost, a new model is formulated for perishable food delivery problems by electric vehicles (PFDP-EV), which considers vehicle capacity constraints, travel time constraints, time window constraints, and so on. An improved particle swarm optimization with particle refactor operator (IPSO-PRO) is developed to solve the proposed model. For the IPSO-PRO, a particle refactor operator is designed to help reconstruct the unqualified particles, and an elite selection strategy and an adaptive weighted strategy are used to improve the performance. Then, extensive efforts are conducted to verify the proposed method. First, the parameters of IPSO-PRO are tuned based on the Taguchi method. Second, small-scale, medium-scale, and large-scale perishable food delivery instances (19 instances) are simulated to evaluate the performance, and the results show that IPSO-PRO achieves the best average gap of 0%. Finally, based on a simulation case, the result and sensitivity analysis are conducted to reveal insightful management insights, which provides decision support for perishable food delivery problems.