Novel composition test functions algorithm for numerical optimization

摘要

Since coming out, novel composition test functions have received wide attention from evolutionary computation researchers and have now become the target functions for numerical optimization algorithms. However, its numerical optimization can be transformed into numerical optimization of one-dimensional functions, which significantly reduces optimization level of difficulty. A novel composition test functions algorithm for numerical optimization is proposed, which quotes a muti-population coevolutionary algorithm for numerical optimization and uses it to optimize the one-dimensional functions. The experiments proved the algorithm for numerical optimization of novel composition test functions converges to the global optimal solutions.