Exact methods in optimum disassembly sequence search for problems subject to sequence dependent costs

• Published in: Omega (Oxford) Vol. 34; no. 6; pp. 538 - 549
• Main Author: Lambert, A.J.D.
• Format: Journal Article
• Language: English
• Published: Oxford Elsevier Ltd 01.12.2006; Elsevier; Pergamon Press Inc
• Series: Omega
• Subjects: Assembly planning; Assembly planning Disassembly planning Mathematical programming Network Traveling salesperson problem Optimization; Disassembly planning; Linear programming; Mathematical models; Mathematical programming; Network; Network flow problem; Optimization; Studies; Traveling salesman problem; Traveling salesperson problem; Disassembly planning; Traveling salesperson problem; Optimization; Assembly planning; Network; Mathematical programming
• ISSN: 0305-0483; 1873-5274
• DOI: 10.1016/j.omega.2005.01.005

Disassembling complex products is formally approached via network representation and subsequent mathematical modeling, aimed at selecting a good or optimum sequence of disassembly operations. This is done via heuristics, metaheuristics or mathematical programming. In contrast with heuristics and metaheuristics, which select a near-optimum solution, mathematical programming guarantees the selection of the global optimum. This problem is relatively simple if the disassembly costs can be assumed sequence independent. In practice, however, sequence dependent disassembly costs are frequently encountered, which causes NP-completeness of the problem. Although methods, e.g., based on the two-commodity network flow approach, are available to solve this constrained asymmetric Traveling Salesperson problem rigorously, this requires the introduction of integer variables. In this paper, a modification of the two-commodity network flow approach is proposed, which reduces the number of integer variables. This is applied to product structures that can be represented by a disassembly precedence graph. It is demonstrated that use of integer variables is completely avoided by iteratively solving a binary integer linear programming problem. This appears to be more efficient than solving the corresponding integer linear programming problem. It is demonstrated, on the basis of some cases, that this method might provide the exact solution of problems with increased complexity compared to those discussed so far in the literature. This appears particularly useful for evaluating heuristic and metaheuristic approaches.