Study of initial population in evolutionary programming

This paper presents a method of generating the initial population of evolutionary programming (EP) for continuous global optimization by using a quasi-random sequence (QRS). instead of a pseudo-random sequence (PRS). The initialized population generated is much more evenly distributed. which can avoid causing rapid clustering around an arbitrary local optimum, and leads to a more reliable solution. The expected number of generations to find the global optimum by EP based on a QRS initialized population is shown theoretically to be less than that based on a PRS initialization population. Numerical test results confirm this advantage.