Applied Intelligence

, Volume 18, Issue 1, pp 105–111

Ant Colony Optimization with Global Pheromone Evaluation for Scheduling a Single Machine

  • Daniel Merkle
  • Martin Middendorf
Article

DOI: 10.1023/A:1020999407672

Cite this article as:
Merkle, D. & Middendorf, M. Applied Intelligence (2003) 18: 105. doi:10.1023/A:1020999407672

Abstract

Ant Colony Optimization (ACO) is a metaheuristic that has recently been applied to scheduling problems. We propose an ACO algorithm for the Single Machine Total Weighted Tardiness Problem and compare it to an existing ACO algorithm for the unweighted problem. The proposed algorithm has some novel properties that are of general interest for ACO optimization. A main novelty is that the ants are guided on their way through the decision space by global pheromone information instead of using only local pheromone information. It is also shown that the ACO optimization behaviour can be improved when priority scheduling heuristics are adapted so that they appropriately reflect absolute quality differences between the alternatives before they are used by the ants. Further improvements can be obtained by identifying situations where the ants can perform optimal decisions.

ant algorithms scheduling tardiness 

Copyright information

© Kluwer Academic Publishers 2003

Authors and Affiliations

  • Daniel Merkle
    • 1
  • Martin Middendorf
    • 2
  1. 1.Institute AIFBUniversity of KarlsruheKarlsruheGermany
  2. 2.Department of Computer ScienceUniversity of LeipzigLeipzigGermany

Personalised recommendations