Comparison of Performance between Genetic Algorithm and Breeding Algorithm for Global Optimization of Continuous Functions

This paper indicates a practical way and conditions for the algorithms to achieve global optimization according to its probability characteristic. Based on this, the convergence performances of conventional genetic algorithm (GA) and breeding algorithm (BA) are estimated and compared according to the globability, accuracy and computation cost. The results show that the conventional GA can not perform not only effective global search but also the accurate local search. For the same probability of global optimization, BA can achieve more accurate computation at about half cost of that of conventional GA. Furthermore, the computation accuracy of BA can be controlled by the length of binary strings. This study reveals the pitfalls existing in conventional GA and designates a reasonable direction for the choice and improvement of the strategies of global optimization.