Optimization of distributed systems' fusion based on Climb-ant algorithm

This paper proposes an optimization algorithm which combines climb algorithm and ant algorithm and use it to solve optimization problems of fusion rules in distributed detection systems.Firstly, algorithm takes Bayes risk as the cost function, uses Climb-ant algorithm to optimize the sensors' decision threshold and Bayes criteria to solve the fusion rules of fusion center. Ultimately, we get the system's optimal (or suboptimal) fusion rules and realize the optimization of fusion rules in distributed detection system. Theoretical analysis and experimental results show that the new method has higher accuracy and efficiency of the solution than the traditional MAP (maximum a posteriori), SA-MAP, GA-MAP, and Coordinate Rotation Method.