Multi-agent optimization model and algorithm for perishable food location-routing problem with conflict and coordination

Considering high distribution cost and high product loss, and complicated relations between a fresh food seller and a hired transportation company, a new model is formulated for perishable food locationrouting problem with conflict and coordination (PFLRP-CC). In this model, customers' time windows are taken as fuzzy variables, the seller as the leader aims at minimizing total costs, while the transportation company is the follower only caring about transportation costs. A GAPSO hybrid algorithm is developed to solve the PFLRP-CC, where an elite selection and an adaptive weighted particle optimization are adopted. Taguchi analysis is used to obtain reasonable values for GAPSO parameters. Small-size problems are solved by GAPSO and then compared with the exact method using CPLEX, the results of which show that GAPSO reduced the computing time by 96.17%; As for medium and large-size cases from Barreto and Prins benchmarks, compared with HybridGA and the best known results, the proposed GAPSO can effectively converge to optimal solutions for medium cases, and find approximate optimal solutions for large cases, which indicate that the GAPSO is efficient and effective for solving the real PFLRP-CC.