Multi-group Flower Pollination Algorithm Based on Novel Communication Strategies

Abstract

Multi-group Flower Pollination Algorithm (MFPA) based on novel communication strategies was proposed with an eye to the disadvantages of the Flower Pollination Algorithm (FPA), such as tardy convergence rate, inferior search accuracy, and strong local optimum. By introducing a parallel operation to divide the population into some groups, the global search capability of the algorithm was improved. Then three new communication strategies were proposed. Strategy 1 combined high-quality pollens of each group for evolution and replaced the old pollens. Strategy 2 let each group’s inferior pollens approaching to the optimal pollen. Strategy 3 was a combination of strategies 1 and 2. Then, experiments on 25 classical test functions show that MFPA based on novel communication strategies has a good global optimization ability, improving the convergence speed and accuracy of the FPA. Thus, we compare MFPA using three strategies with FPA and PSO, its result shows that MFPA is better than FPA and PSO. Finally, we also applied it to two practical problems and achieved a better convergence effect than FPA.