"A GRASP metaheuristic for a districting problem arising in urban distribution"
Andrés Tejada, Olga Usuga and Juan G. Villegas R.

This paper presents a simple greedy randomized adaptive search procedure (GRASP) embedded in a decision support system for the design of delivery districts in urban distribution. The pro-posed GRASP aims at balancing the workload of delivery routes by considering both the travel time within customers and the service time at the customers. To compute travel times we use a continuous approximation of the length of a travelling salesman tour for each district. Computational experiments, with the data of a major Colombian company serving hundreds of stores a day, reveal a potential reduction of the workload imbalance of their routes by more than 50%.