Novel binary differential evolution algorithm based on Taper-shaped transfer functions for binary optimization problems

Abstract

In order to efficiently solve the binary optimization problems by using differential evolution (DE), a class of new transfer functions, Taper-shaped transfer function, is firstly proposed by using power functions. Then, the novel binary differential evolution algorithm based on Taper-shaped transfer functions (T-NBDE) is proposed. T-NBDE transforms a real vector representing the individual encoding into a binary vector by using the Taper-shaped transfer function, which is suitable for solving binary optimization problems. For verifying the practicability of Taper-shaped transfer functions and the excellent performance of T-NBDE, T-NBDE is firstly compared with binary DE based on S-shaped, U-shaped and V-shaped transfer functions, respectively. Subsequently, it is compared with the state-of-the-art algorithms for solving the knapsack problem with a single continuous variable (KPC) and the uncapacitated facility location problem (UFLP). The comparison results show that Taper-shaped transfer functions are competitive than existing transfer functions, and T-NBDE is more effective than existing algorithms for solving KPC problem and UFLP problem.