A COMPARISON OF THREE OPTIMIZATION METHODS FOR SCHEDULING MAINTENANCE OF HIGH COST, LONG-LI.pdf

Proceedings of the 2002 Winter Simulation Conference E. Yücesan, C.-H. Chen, J. L. Snowdon, and J. M. Charnes, eds. A COMPARISON OF THREE OPTIMIZATION METHODS FOR SCHEDULING MAINTENANCE OF HIGH COST, LONG-LIVED CAPITAL ASSETS Terry M. Helm Steve W. Painter W. Robert Oakes Los Alamos National Laboratory Los Alamos, NM 87545, U.S.A. ABSTRACT A range of minimization methods exist enabling planners to tackle tough scheduling problems. We compare three scheduling techniques representative of “old” or standard technologies, evolving technologies, and advanced tech- nologies. The problem we address includes the complica- tions of scheduling long-term upgrades and refurbishments essential to maintaining expensive capital assets. We con- centrate on the costs of being able to do maintenance work. Using a standard technology as the baseline technique, Constraint Programming (CP) produces a 50-yr mainte- nance approach that is 31% less costly. Genetic Program- ming produces an approach that is 60% less costly. 1 INTRODUCTION Simulation “is an established method” for assessing per- formance capabilities and operational domains for a wide range of industries (Splanemann 2001). Production and production-related operations planners often turn to simu- lation technologies to assist operators in dealing with com- plex production requirements particularly in highly con- strained environments (Wang and Handschin 1999). Simulation technologies permit the planner/engineer to quickly assess numerous operational alternatives represent- ing useful solution domains that otherwise cannot be “seen” (Contaxis et al 2000; Bretthauer et al 1998). One of the most critical production-related decisions facing industrial planners is when to do what. It is well recognized, for example, that production effectiveness is symbiotically dependent on timely maintenance as well as on-time product delivery (Kelleher 1997). The ubiquitous ent