The Winner Determination Approach of Combinatorial Auctions based on Double Layer Orthogonal Multi-Agent Genetic Algorithm

The selection of winning bids in the first-price seal combinatorial auction is an extremely complex problem, and in fact it has been shown to be NP-complete. The meta-heuristic algorithms put the problem to be solved into a proper science perspective. It proposes a winner determination approach of combinatorial auctions based on double layer orthogonal multi-agent genetic algorithm in this paper. In this proposed algorithm, the lower-layer Orthogonal Multi-Agent Genetic Algorithm (OMAGA) is applied to searching the optimal solution of the giving combinatorial auctions optimization problem; the upper-layer OMAGA is used for optimizing the parameter of lower-layer OMAGA. The optimization performance of this proposed approach has been improved largely by efficaciously integrating the orthogonal design, multi-agent technology, genetic algorithm and parameter optimization into together. Simulation experiment result suggests that this proposed approach is feasible, correct and valid.