Abstract
Providing a satisfying delivery service is an important way to maintain the customers’ loyalty and further expand profits for manufacturers and logistics providers. Considering customers’ preferences for time windows, a bi-objective time window assignment vehicle routing problem has been introduced to maximize the total customers’ satisfaction level for assigned time windows and minimize the expected delivery cost. The paper designs a hybrid multi-objective genetic algorithm for the problem that incorporates modified stochastic nearest neighbour and insertion-based local search. Computational results show the positive effect of the hybridization and satisfactory performance of the metaheuristics. Moreover, the impacts of three characteristics are analysed including customer distribution, the number of preferred time windows per customer and customers’ preference type for time windows. Finally, one of its extended problems, the bi-objective time window assignment vehicle routing problem with time-dependent travel times has been primarily studied.
Highlights
In many distribution networks, the deliveries are made within a scheduled time window because many operational processes such as inventory management and scheduling of personnel heavily depend on it, especially in retail [1,2]
The computation time of SNNIN20 for the instances with 15 customers is on average about 74 s at a run. These results clearly indicate that compared with no local search (NLS), stochastic nearest neighbour (SNN), and interval of 10 generations (IN10), our hybrid multi-objective genetic algorithm (HMOGA), SNNIN20, is able to better approximate the Pareto-optimal sets of the instances within a reasonable time
The average deviation between the minimal expected delivery costs and their optimal values is 4%, which is satisfactory considering the huge scale of variables, more than NT $ NC + NC2 $ NS + NV $ NC $ NS
Summary
The deliveries are made within a scheduled time window because many operational processes such as inventory management and scheduling of personnel heavily depend on it, especially in retail [1,2]. The problem, called the time window assignment vehicle routing problem, was firstly introduced by Spliet and Gabor [1], who studied how to design one time window that has a predetermined width for each customer from continuous exogenous time windows so as to minimize the expected delivery cost. To maintain the customers’ loyalty and further expand the profits for manufacturers and logistics providers, a bi-objective time window assignment vehicle routing problem (BOTWAVRP) is introduced to maximize the total customers’ satisfaction level for the assigned time windows and to minimize the expected delivery cost [3]. This paper designs a hybrid multi-objective genetic algorithm (HMOGA) for the problem. On this basis, the impacts of three problem characteristics are deeply analysed.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
