Approximation algorithm for multi-criterion nonlinear 0-1 programming based on entropy matrix

In order to simplify the solving of two-class matching problem under multiple criterions, a model of multi-criterion nonlinear 0-1 programming is established, in which constraints on variables are transferred to objective function to reduce the difficulty in solving the problem. A greedy approximate algorithm based on entropy matrix computing is designed to solve this nonlinear 0-1 programming model more efficiently. By calculating the value of entropy which indicates the matching degree of two types of objects, the algorithm can optimize the scheduling matching sequence, then the solution calculated by the algorithm can approximate the optimal solution more quickly. Simulation results show that the algorithm proposed in this paper is superior to other approximate algorithms for solving single-criterion nonlinear 0-1 programming problem. For multi-criterion nonlinear 0-1 programming problem, the approximate solution given by this algorithm is more close to the optimal solution while the computing time is decreased exponentially.