Meta Heuristic Algorithms for Vehicle Routing Problem with Stochastic Demands

Abstract

Problem statement: The shipment of goods from manufacturer to the consumer is a focal point of distribution logistics. In reality, the demand of consumers is not known a priori. This kind of distribution is dealt by Stochastic Vehicle Routing Problem (SVRP) which is a NP-hard problem. In this proposed work, VRP with stochastic demand is considered. A probability distribution is considered as a random variable for stochastic demand of a customer. Approach: In this study, VRPSD is resolved using Meta heuristic algorithms such as Genetic Algorithm (GA), Particle Swarm Optimization (PSO) and Hybrid PSO (HPSO). Dynamic Programming (DP) is used to find the expected cost of each route generated by GA, PSO and HPSO. Results: The objective is to minimize the total expected cost of a priori route. The fitness value of a priori route is calculated using DP. In proposed HPSO, the initial particles are generated based Nearest Neighbor Heuristic (NNH). Elitism is used in HPSO for updating the particles. The algorithm is implemented using MATLAB7.0 and tested with problems having different number of customers. The results obtained are competitive in terms of execution time and memory usage. Conclusion: The computational time is reduced as polynomial time as O(nKQ) time and the memory required is O(nQ). The ANOVA test is performed to compare the proposed HPSO with other heuristic algorithms.