Projects

This study considered a distribution network of a brewery manufacturer delivering products to retail outlets in different geographical locations with distinct demand and operating time windows. The company possesses a heterogeneous fixed fleet of trucks in four different capacities. An increased distribution cost and the underutilization of resources are major issues of the company due to inefficient distribution planning method. This study aims to generate efficient delivery routes with minimum travelled distance assigning a set of outlets to utilize more than 80% of the truck capacity.

This problem was formulated as a Vehicle Routing Problem (VRP) with the heterogeneous fleet, which is an integer programming problem belonging to the combinatorial optimization category. The model was solved using both Simulated Annealing (SA) and Genetic Algorithms (GA). Initially, the optimal sequence of all outlets in terms of distance was generated using a greedy approach. Trucks were then assigned to the best set of outlets achieving the required capacity utilization levels given the minimum travel distance achieved. Then statistically compared the results obtained from two algorithms and selected the best approach. Then scenario analysis based on fleet compositions is conducted and statistically identified better fleet composition to address the problem. The analysis revealed that by using genetic algorithm and excluding the lowest capacitated truck from the fleet gave the best results. The total distance reduced by 13% compared to existing system and average utilization is improved to 96%