Parameter’s setting of the ant colony algorithm applied in preventive maintenance optimization

Meta-heuristic applications have evolved a lot nowadays and have been used in many domains. However, their parameter setting stills, till now, a serious problem which influences their efficiency and their attitude. Under the title of preventive maintenance optimization, the Ant Colony (AC) is used as a search technique for the best intervention dates to preventively maintain a series-parallel system. The AC proves once again its efficiency (Samrout, M., Yalaoui, F., Châtelet, E., & Chebbo, N. (2005). Reliability Engineering & System Safety, 89, 346–354), but the parameter setting’s issue persists as a critical issue which needs more analysis. This paper offers a detailed study about the influence of the parameters on the AC operation. An experimental plan design is done. Obtained results are coherent with the convergence conditions of the ant colony algorithms. Based on those latter, improvements are done to the original ant colony algorithm and interesting results are obtained.