Discrete greedy flower pollination algorithm for spherical traveling salesman problem

This paper deals with the spherical traveling salesman problem. In this problem, all cities are located on the surface of a sphere and the cities must be visited exactly once in a tour. We propose a new and effective meta-heuristic algorithm with greedy behavior for solving this problem. The proposed algorithm is based on the discrete flower pollination algorithm, which is a bio-inspired meta-heuristic algorithm enhanced by order-based crossover, pollen discarding behavior and partial behaviors. To evaluate the proposed algorithm, it is compared with four effective existing algorithms (the genetic algorithm, two variants of the genetic algorithm and tabu search) on a set of available spherical traveling salesman instances. The results show the superiority of our algorithm in both solution quality and robustness of the solutions.