Task Scheduling Under Gang Constraints

Abstract

In this paper, a short-term manpower planning problem is considered where workers are grouped into gangs to support reliable and efficient operations. The goal is to minimise the total number of workers required by determining an appropriate gang structure, assignment of tasks to gangs, and schedule for each gang. We model such a problem as a multi-mode task scheduling problem with time windows and precedence constraints. While the gang structure and assignment of tasks is optimised by a tabu search heuristic, each gang's schedule is generated by solving the corresponding one-machine scheduling problem by an iterated Schrage heuristic. Because the evaluation of a tabu search move is computationally expensive, we propose a number of ways to estimate a move's impact on the solution quality.