An optimization ant colony algorithm based on fixed point theory

the convergence accuracy and stability of ant colony optimization algorithm (ACO) are in urgent need of improvement ever since it has been proposed. Existing research attempt to optimize the algorithm from various perspectives, but most of them are relatively ex prate or application specific. In this paper, we design an elite strategy to optimize the initial parameters and build a Fixed-point ACO improved algorithm. The process is that it first converts the problem of targeted function optimization to a problem of fixed-point searching. Then the solution of the equation set is obtained by simplicial algorithm (SA) of fixed-point theory. Finally, the solution set will be regarded as the initial population of ACO algorithm and the remaining parameters are set accordingly. The experimental study is carried with five testing functions from UCI database. The results indicate that FP-ACO algorithm is significantly better than the conventional algorithm both on the average speed of convergence and the average accuracy of global optimum. Besides, the searching process of FP-ACO algorithm has a better stability by presenting a state of continuous optimization.