An Evolutionary Algorithm for Zero-one Nonlinear Optimization Problems Based on an Objective Penalty Function Method

In many evolutionary algorithms, as fitness functions, penalty functions play an important role. In order to solve zero-one nonlinear optimization problems, a new objective penalty function is defined in this paper and some of its properties for solving integer nonlinear optimization problems are given. Based on the objective penalty function, an algorithm with global convergence for integer nonlinear optimization problems is proposed in theory. As a further application of the objective penalty function, a simple novel evolutionary algorithm is presented for solving zero-one nonlinear optimization problems. Numerical results on several examples show that the proposed evolutionary algorithm seems effective and efficient for some zero-one nonlinear optimization problems.