An Improved Cultural Algorithm for Scheduling of Precedent Tasks

In order to efficiently process some tasks with precedent constraints, a cultural algorithm based on greedy stochastic process is proposed. The improved cultural algorithm firstly gives an initial population with obvious features, which provides a precondition for accelerating the convergence speed of cultural algorithm. The cultural algorithm introduces the belief space, in which some excellent features produced by generations of individuals are stored. The culture formed by these features and the group space can influence each other, thereby improving the convergence speed and the overall optimization ability. The final simulation results show that the cultural algorithm based on greedy stochastic process can obtain a relatively better scheduling scheme in an acceptable time.