Performance assessment of foraging algorithms vs. evolutionary algorithms Mohammed El-Abd Computer Engineering Department, American University of Kuwait, P.O. Box 3323, Safat 13034, Kuwait article info Article history: Received 14 December 2009 Received in revised form 1 September 2011 Accepted 5 September 2011 Available online 16 September 2011 Keywords: Foraging algorithms Evolutionary algorithms Performance comparison Evolutionary optimization abstract The class of foraging algorithms is a relatively new field based on mimicking the foraging behavior of animals, insects, birds or fish in order to develop efficient optimization algo- rithms. The artificial bee colony (ABC), the bees algorithm (BA), ant colony optimization (ACO), and bacterial foraging optimization algorithms (BFOA) are examples of this class to name a few. This work provides a complete performance assessment of the four men- tioned algorithms in comparison to the widely known differential evolution (DE), genetic algorithms (GAs), harmony search (HS), and particle swarm optimization (PSO) algorithms when applied to the problem of unconstrained nonlinear continuous function optimiza- tion. To the best of our knowledge, most of the work conducted so far using foraging algo- rithms has been tested on classical functions. This work provides the comparison using the well-known CEC05 benchmark functions based on the solution reached, the success rate, and the performance rate. Ó 2011 Elsevier Inc. All rights reserved. 1. Introduction The foraging behavior of any living organism is defined as how this organism behaves in order to locate, handle, and in- gest food. The approach taken in this behavior is referred to as the search strategy. For many animals, this strategy is usually broken down to three steps, namely: searching for and locating the prey, attacking the prey, and ingesting the prey [28]. The relative importance of these steps depends on the animal type and the environment characteristics. Search strategies adopted by many living organisms inspired the development of different optimization algorithms cur- rently adopted for various engineering applications. The main idea was for the search agents to imitate the foraging behavior of organisms in order to search for a solution of the problem. The algorithms included ABC [14] and BA [30] mimicking the foraging behavior of honey bees. BFOA [28], which mimics the foraging behavior of a swarm of E.coli bacteria, and ACO [10], which mimics the foraging behavior of ants. Previous studies performed to assess the performance of some of these algorithms included the work in [20] showing that ABC performs better than PSO, an evolutionary algorithm (EA), and DE on a small suite of classical benchmark functions. The study in [17] compared ABC against PSO, a genetic algorithm (GA), DE, and an evolutionary strategy (ES) algorithm on a lar- ger number of functions. It was shown that the performance of ABC is better than or at least similar to those algorithms while having a smaller number of parameters to tune. The work in [16] compared ABC to HS, and BA. The comparison was based on a small set of classical functions and ABC showed superior performance over both algorithms while producing reasonable results for higher dimensions. The aforementioned studies suffer from two limitations. First, these studies only compare a certain algorithm to either evolutionary algorithms or other foraging algorithms but there is no single study that covers both. Second, the comparisons are based on a set of more or less classical functions. The proposed study overcomes these drawbacks by assessing the per- 0020-0255/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.ins.2011.09.005 E-mail address: melabd@auk.edu.kw Information Sciences 182 (2012) 243–263 Contents lists available at SciVerse ScienceDirect Information Sciences journal homepage: www.elsevier.com/locate/ins