Soccer Game Optimization: An Innovative Integration of Evolutionary Algorithm and Swarm Intelligence Algorithm

Introduction

In the last two decades, metaheuristic algorithm has emerged as powerful optimization algorithm, especially in engineering, science, business, economic and finance. The algorithm provides a higher framework of heuristic methods to explore the search space more efficiently and effectively. The term metaheuristics was introduced by Glover (1986). The word derives from the Greek words: heuriskein and meta. Heuriskein means “discover” or “find” and it is generally accepted as the term for approximate methods while meta means “in upper level”. Metaheuristic can be described as a higher level of general methodology that provides guidance based on the heuristic search to solve various optimization problems. An important property that distinguishes between heuristic and metaheuristic algorithms is: the heuristic algorithm is problem dependent while the metaheuristics algorithm is problem independent.

A lot of successful applications of metaheuristic algorithms have been published. The metaheuristic algorithms emerge as a powerful method to solve various optimization problems (Bianchi et al., 2008; Deep et al., 2009). The method gains significant interests in research and industrial practices due to its effectiveness and general applicability. In many cases, the classical approaches based on mathematical and dynamic programming are feasible only for small size instances of problems and generally require a lot of computational efforts, therefore the metaheuristics turn up into promising alternatives to classical optimization methods (Bianchi et al., 2008). The metaheuristic algorithms exploit randomness and set of rules to produce solutions. Due to the nature of heuristic methods, the metaheuristic methods do not guarantee optimal global solutions; however, they will provide acceptable solutions for large-sized and complex problems in a reasonable time (Talbi, 2009). Compared to the exact solution, metaheuristic algorithms are more flexible in their adaptability to fit the need of various optimization problems and they do not need to put formulation of the optimization problem. However, the algorithms need considerable problem specific adaptation in order to achieve good performance (Sorensen and Glover).

Metaheuristic methods have been applied in various problems such as: logistic (Shimizu et al., 2007; Wang and Hsu., 2010,), scheduling (Zhang et al., 2011; Deng and Lin., 2011), data mining (Sorensen and Janssens, 2003; Srinivasa et al. 2007), assembly lines (Lee et al., 2001; Simaria and Vilarinho, 2009; Kim et al., 2000; Purnomo et al., 2011) and supply chain management (Silva et al., 2009; Melo et al., 2012). The development of metaheuristic methods continually evolves to cope with the growing of optimization problems. Therefore, developing a new meta-heuristic method is an important contribution to solve optimization problems in engineering and operations research. The objective of this chapter is to describe a new metaheuristic method, which combines the evolutionary algorithm and swarm intelligence concepts. The method mimics the soccer player movement during a game, and for this rationale, we name it Soccer Game Optimization (SGO). The proposed method combines the local search and global search with the use of cognitive learning as well as social learning. This chapter illustrates the work of the proposed method. Examples of continuous and discrete problems are provided in order to help understand the proposed method easily.