Generating disassembly sequences using exact and heuristic methods applied to disassembly precedence graphs

This paper discusses the search for (sub)optimum disassembly sequences for products that can be represented by disassembly precedence graphs and that are subjected to sequence-dependent disassembly costs. A new exact approach, based on iteratively solving a binary linear programming problem, is applied. This method appears more efficient than standard integer linear approaches. Its application to medium sized problems, which are frequently encountered in practice, is demonstrated. Besides this, the iterative method is valuable as a benchmark for validation of heuristic algorithms. This is demonstrated by application to a new heuristic. Due to the benchmark, this heuristic is adapted such that appropriate convergence of the heuristically found solution to the optimum is guaranteed