A discrete teaching–learning based optimization algorithm with local search for rescue task allocation and scheduling

Abstract

The allocation and scheduling of the emergency rescue forces is a fundamental task in emergency management. This paper aims to address the allocation and scheduling problem to minimize the average completion time of all rescue teams by using a discrete teaching–learning based optimization algorithm with local search (DTOLS). First, an improved k-means clustering algorithm with constraints is proposed to assign tasks to rescue teams based on the location of rescue tasks. Second, a hybrid discrete optimization algorithm based on a teaching–learning mechanism is designed to generate the task scheduling sequence for each rescue team as an initial solution. Next, an efficient two-phase local search strategy is presented to improve the current solution. For three neighborhood task moves based on problem characteristics, which contains insert task within a team, swap tasks within a team, insert task between teams, the speed-up techniques are introduced to reduce the computational complexity of calculating completion time of a rescue team. Finally, the parameters of DTOLS are calibrated by Taguchi method to determine appropriate values. DTOLS is compared with the state-of-the-art algorithms, and the experimental results demonstrate the effectiveness of DTOLS in solving a set of test instances.