This paper aims to study the vehicle routing problem with dynamic customers considering dual service (including home delivery [HD] and customer pickup [CP]) in the last mile delivery in which three decisions have to be made: determine routes that lie along the HD points and CP facilities; optimize routes in real time, which mode is better between simultaneous dual service (SDS, HD points and CP facilities are served simultaneously by the same vehicle); and respective dual service (RDS, HD points and CP facilities are served by different vehicles)?Design/methodology/approachThis paper establishes a mixed integer linear programing model for the dynamic vehicle routing problem considering simultaneous dual services (DVRP-SDS). To increase the practical usefulness and solve large instances, the authors designed a two-phase matheuristic including construction-improvement heuristics to solve the deterministic model and dynamic programing to adjust routes to dynamic customers.FindingsThe computational experiments show that the CP facilities offer greater flexibility for adjusting routes to dynamic customers and that the SDS delivery system outperforms the RDS delivery system in terms of cost and number of vehicles used.Practical implicationsThe results provide managerial insights for express enterprises from the perspective of operation research to make decisions.Originality/valueThis paper is among the first papers to study the DVRP-SDS. Moreover, this paper guides the managers to select better delivery mode in the last mile delivery.
Kybernetes – Emerald Publishing
Published: Apr 8, 2020
Keywords: Dynamic programming; Vehicle routing problem; Simultaneous dual services; Dynamic customers; Hybrid heuristics