An Improved Bat Algorithm for Solving Nonlinear Mixed Integer Optimization Problems

This paper proposed an improved BA algorithm for solving nonlinear mixed integer optimization problems(NMIP) problem, An adaptive local search step size is constructed, and the integer variable is searched by unit step size, this improves the local search ability of the improved algorithm. Natural selection principle can improve the global search ability of the improved algorithm. The improved algorithm uses a feasible solution to ensure the correct search direction of the algorithm. Experimental results of sixteen test function show that the improved bat algorithm is not inferior to the improved particle swarm optimization algorithm for solving nonlinear mixed integer optimization problems.