An Insect-Based Algorithm for the Dynamic Task Allocation Problem

An insect-based algorithm inspired by the division of labor in insect colonies is proposed and applied to the solution of an online scheduling problem. A painting facility is considered for illustrating the problem: Trucks leave an assembly line to get painted in painting booths. The goal is to minimize the makespan, that is, the time needed for painting all given trucks. In this paper we address two issues. First, we propose and analyze four modifications of an insect-based algorithm previously introduced by Cicirello and Smith. Second, we propose an extension of the dynamic threshold model of Theraulaz et al. that was originally used for handling the case in which all booths are identical; in our extension we address the heterogeneous case in which different booths might require different processing times for completing a same task. The paper contains an empirical comparison of different insect-based algorithms for the problem discussed.