Rank-Based Version of Ant System in production scheduling

Abstract

The paper presents the application of the Rank-Based Version of Ant System algorithm to a common production scheduling problem. This problem is Flow Shop Scheduling. The paper presents the results using the Taillard benchmark dataset. The Rank-Based Version of Ant System (RBAS) is an improved version of the original Ant System (AS) algorithm. The goal of RBAS is to improve the convergence of the original algorithm. The algorithm uses rank-based pheromone updating. In this method, not all ants contribute equally to the pheromone deposit. Ants that find the best solutions have a greater influence on the search process. Ants that find weaker solutions have a smaller impact on the creation of new solutions. As a result, the algorithm can have faster convergence and provide high-quality solutions. The paper presents the Flow Shop Scheduling, the Rank-Based Version of Ant System, and the running results on the Taillard benchmark dataset.