Article Preview
Top1. Introduction
Artificial Bee Colony (ABC) algorithm proposed by Karaboga (Karaboga, 2005) is a new swarm intelligence based metaheuristic algorithm that simulates the natural behavior of a honey bees swarm when they searching for food sources. In this algorithm, the artificial bee colony consists of three sets of bees: employed bees, onlooker bees, and scout bees.
Artificial Bee Colony algorithm firstly, initializes randomly distributed population of size SN. Where SN represents the number of positions of the food sources and each one represents a possible solution to the optimization problem. For every food source, there is only one employed bee, so the number of solutions is equal to the number of employed bees or the onlooker bees. Each solution is a D dimensional vector, where D is the number of optimization parameters (Karaboga & Akay, 2009). The quality for each solution, which is evaluated by a fitness function, represents the amount of nectar for the corresponding food source. After initialization, when a set of food sources is selected randomly by the bees and evaluated their quality, there are three basic steps repeated a predetermined number of cycles called MCN where MCN is the Maximum Cycle Number, or until a termination criterion is satisfied.
Choosing a random value that attempts to balance the exploitation and exploration phase in this algorithm will affect on its performance of searching process. Also, ABC algorithm proved its superiority over other algorithms when it’s applied for both benchmark functions and real world problems. Hybridizing the ABC algorithm with operators applied on Genetic algorithm may enhance its performance.
Genetic Algorithm (GA) is a randomized search procedure based on the principles of natural selection and natural genetics (Mitchell, 1998) (Goldberg, Deb, & Korb,1989). This algorithm involves simple principles: selection, crossover, and mutation. Among the GA operators is the crossover in which two individuals (parents) are selected randomly from the population to generate new individuals (off- springs) that share some attributes with each parent. There are several crossover techniques, and the performance of genetic algorithm is influenced by the used crossover technique (Durand & Alliot, 1998) (Kaya, Uyar,& Tekin, (2011),(Spears & Anand, 1991).
There are several advantages for the crossover operator, it maintains the population diversity by generating new chromosomes (individuals) that may be better than their parents, so it avoids the premature convergence (Goldberg, Deb, & Korb,1989) (Mitchell, 1998). Also, the crossover operator represents the key to the strength of genetic algorithm (Spears & Anand, 1991), (Chen & Smith, 1999). In that genetic algorithm with crossover operator outperforms a genetic algorithm that does not apply crossover during the evolution process. This operator considered as the main search operator in genetic algorithm since it exploits the search space information from the individuals in the population. In addition, crossover ensures preserving of common alleles. In particular, the multi-parent crossover is not new in the literature. Many studies proved that using genetic algorithm with multi-parent crossover operator outperformed using two parents crossover operator (Elsayed, Sarker, & Essam, 2011)(Eiben, 1997) (Eiben, Kemenade, Eiben & Kok,1995) (Deb, Joshi, & Anand, 2002).