202 An Approach for Line Configuration Based on Optimal Bipartite Graph Matching

A team's line-up plays a critical role in soccer. Motivated by the theory of the optimal matching in bipartite graph, we propose an approach based on Kuhn-Munkres algorithm for line-up configuration according to the specific formation and team information. A weighted bipartite graph is constructed to model the relationship between players and positions: there are two kinds of vertex sets, one represents players, and the other represents positions. The weight of every edge between two sets represents the score for some player in the corresponding position. Then Kuhn-Munkres algorithm is employed to compute the maximum weight of a constructed bipartite graph, which is regarded as optimal result. Besides by prohibiting some of the players playing and then re-scheduling the algorithm we find a number of good line-up. Next through data mining and decision tree algorithm we get the player features of best line-up in the corresponding positions, and compare them with the result. Theoretical analysis and experimental results show that the method proposed can complete the best line-up configuration function, and as a reference in the non-professional football.