"Planning of ‘last-mile’ delivery for a Colombian dairy company using a biased-randomized multi-start algorithm"
Carlos L. Quintero-Araujo, Carlos Vega-Mejía and Andrés Muñoz-Villamizar

This document focuses on the ‘last-mile’ distribution of dairy products of a company in Colombia. To improve the current operational and economic performance of the company a heuristic solution method is proposed to solve a Vehicle Routing Problem with Time Windows. This problem characterizes the ‘last-mile’ delivery problem that the dairy company faces on its day-to-day operation. The proposed heuristic method consists of a multi-start algorithm with biased randomization and is validated utilizing historical data from the Company. Computational results show that, employing the proposed heuristic method, a more accurate planning of delivery routes can be achieved in a shorter processing time than with the current manual procedure the company us-es. The proposed solution procedure increases the percentages of space utilization of the vehicles required for the ‘last-mile’ delivery task. This increase represents monetary savings for the Company.