Virtual enterprises partner selection based on reverse auctions

Partner selection is a key issue in the development of an effective coalition formation mechanism for virtual enterprises (VE). Combinatorial reverse auction can be applied by a firm to select the best partners to minimize the cost in forming VE. The objectives of this paper are to propose architecture for selecting partners based on combinatorial reverse auction mechanism to minimize the cost of VE, develop algorithms to find a near-optimal solution efficiently, and implement a prototype system based on the proposed algorithms. We formulate the partner selection problem based on combinatorial reverse auctions and apply Lagrangian relaxation technique to solve the problem. Our partner selection solution algorithms include an algorithm for solving bidders' subproblems by exploiting their problem structures, a subgradient algorithm for solving the dual problem, and a heuristic algorithm for finding a near-optimal solution. In addition to theoretical development, we also implement a prototype system based on the proposed algorithms and web services technologies to verify the effectiveness of our methodology.