A real-world case study of a vehicle routing problem under uncertain demand

Abstract

The research scope of the real-world logistics industry case study is extended by taking uncertainty in customer demand into account. The particular vehicle routing planning parameters of the logistics provider under study are formulated and are used in two algorithms. The algorithms solve practical problem cases considering a limited number of drivers and a limited company’s fleet size but unlimited when considering outsourcing. All trucks are allowed to service multiple trips. The computation is based on real-life data sets. The analysis of the running time and the total transportation cost are compared among three competitive methods. The methods are: the technique based on the company’s know-how, a genetic algorithm hybridized with three search operators, and a deterministic annealing hybridized with three search operators. The developed schemes have been proven successful to obtain a near-optimal solution within a reasonable running time. Furthermore, the adaptation of the minimax concept is embedded into the algorithms to find a robust solution for the worst case scenario subject to handling fluctuating situations in demand. In the last phase, two indicators comprising the extra cost and the unmet demand ratios are proposed to help a decision maker to obtain a better view on his decision.