Given a graph G=(V,E) with edge weights we∈ℝ, the optimum cooperation problem consists in determining a partition of the graph that maximizes the sum of weights of the edges with nodes in the same class plus the number of the classes of the partition. The problem is also known in the literature as the optimum attack problem in networks. Furthermore, a relevant physics application exists.
In this work, we present a fast exact algorithm for the optimum cooperation problem. Algorithms known in the literature require |V|−1 minimum cut computations in a corresponding network. By theoretical considerations and appropriately designed heuristics, we considerably reduce the numbers of minimum cut computations that are necessary in practice. We show the effectiveness of our method by presenting results on instances coming from the physics application. Furthermore, we analyze the structure of the optimal solutions.
Optimum attack problem Submodular function minimization Potts glass with many states
Anglès d’Auriac JC, Iglói F, Preissmann M, Sebő A (2002) Optimal cooperation and submodularity for computing Potts’ partition functions with a large number of states. J Phys A, Math Gen 35:6973–6983
Baïou M, Barahona F, Mahjoub R (2000) Separation of partition inequalities. Math Operat Res 25(2):243–254
McCormick ST (2005) Submodular function minimization. In: Aardal K et al. (eds) Discrete optimization. Handbooks in operations research and management science, vol 12. Elsevier, Amsterdam, pp 321–391
Preissmann M, Sebő A (2009) Graphic submodular function minimization: a graphic approach and applications. In: Cook W, Lovász L, Vygen J (eds) Research trends in combinatorial optimization. Springer, Berlin, pp 365–386