Optimal decentralized formation of k-member partnerships (May 9, 2010)

To keep pace with constantly changing markets, many companies are seeking strategic partnerships. In this paper, we assume that a company can electronically provide a profile of the product or service it has to offer. This profile is described in such a way that potential partners can assess the fitness of the company for eventually teaming up. We concentrate on the fully decentralized optimal formation of teams consisting of k members. This problem boils down to developing a decentralized, efficient algorithm for solving a variant of the maximal weighted k-subgraph problem. We provide a first solution, along with an assessment of its performance, thereby concentrating on the feasibility of an actual embedding in real-world scenarios consisting of thousands of companies. In particular, any solution should be highly adaptive when new or fresh information concerning potential partners comes available. Keywords-strategic alliances; product bundling; k-clique matching; variable neighbourhood search; self-stabilization;