Multi-Start Local Search for Sugar Beet Transportation Problem with Equity Regulations for Growers

Abstract

The sugar beet transportation problem under growers' equity regulations is considered. The solution of the problem represents the optimal plan for the delivery of sugar beet collected by several growers at different storage piles to a single sugar mill. The vehicle fleet is homogeneous and vehicles are used several times during the day. Requirements in amount of transported beet must be satisfied for each day of the planning period. Equity regulations are related to fairness among growers, providing the equal opportunities in delivering sugar beet when it is fresh, contains large amount of sucrose, and therefore, is more profitable. Constraints related to equity regulations are formulated as the minimum percentage of beet that must be transported on the harvesting day for each grower. The goal is to minimize the number of used vehicles while maximizing the total sucrose content of the collected sugar beet. The Multi-start Local Search (MLS) metaheuristic is proposed for the considered problem, evaluated on the set of generated instances against CPLEX exact solver and an existing metaheuristic approach based on the Greedy Randomized Adaptive Search Procedure (GRASP). Our experimental results show that MLS outperforms GRASP in average and CPLEX in the case of large-sized instances.