Differential Evolution (DE) is a simple and efficient heuristic algorithm for global optimization over continuous spaces. Because of its outstanding performance, simplicity and remarkable efficiency, DE has being used widely to deal with real-life problems. However DE has some problems such as low precision and prematurity if the scaling factor F is not chosen carefully. In this paper a modified DE algorithm(sFDE) is proposed which uses an adaptive scaling factor sF to replace the invariable factor F in basic DE. This new algorithm performs better than the basic DE over a set of benchmark functions and a practical use for data reconciliation shows its effectiveness also.
展开▼