A New Tabu Search Approach for Optimizing Series-Parallel Multi-State System Performance
Most existing works on redundancy allocation problem (RAP) use traditional binary-state reliability. Binary-state reliability modeling assumes that a system and its elements may experience only two possible states˙. good and failed. The multi-state version of the problem is more recent and has not been suciently studied. To solve multi-state series-parallel systems problem, this paper develops an ecient tabu search algorithm. The system have a range of performance levels from perfect functioning to complete failure. The elements of the system are characterized by their cost, performance and availability. System availability is defined as the ability to satisfy consumer demand. A universal generating function technique is applied to evaluate system availability. The proposed tabu search heuristic determines the minimal cost system configuration under availability constraints. Numerical results are reported for large test problem which is randomly generated. An originality of our approach is that it proceeds by dividing the search space into a set of disjoint subsets, and then applying tabu search to each subspace.
Reliability and availability optimization Meta-heuristics Tabu search Multi-state systems Series-parallel systems Universal generating function
Mohamed OUZINEB Mustapha NOURELFATH Michel GENDREAU
CIRRELT,Universite de Montreal,Departement dinformatique et de recherche operationnelle Montrea CIRRELT,Universite Laval,Departement de genie mecanique,Quebec (Qc.),Canada CIRRELT,Universite de Montreal,Departement dinformatique et de recherche operationnelle Montrea
国际会议
北京
英文
2007-05-30（万方平台首次上网日期，不代表论文的发表时间）