Artificial Bee Colony Algorithm and an Application to Software Defect Prediction

Abstract

The term swarm refers to any restrained collection of interacting agents or individuals. For survival, creatures need to live or perform some tasks collectively such as defending against predators, foraging, mating, etc. An intelligent swarm optimizes a goal or task by responding adaptively to the local and/or global environmental changes in a collective manner. Swarm intelligence research field deals with designing algorithms inspired by the collective behavior of social creatures. Task division and self-organization abilities lead swarm intelligence to occur in a colony and the self-organization is characterized by positive-feedback, negative feedback, fluctuation and multiple interactions in order to use the local information to form a global pattern without a supervision. Ants, termites, birds, and fishes are some examples of social animals that have swarm intelligence and inspire researchers to design problem solving techniques. Bees also are a typical example of creatures performing tasks collectively in nest site selection, nest building, mating, and foraging. Among these activities, foraging might be the most crucial one for the survival of a bee colony. The swarm intelligence in foraging of a honey bee colony inspired Karaboga to design an optimization algorithm, Artificial Bee Colony (ABC), in which the search is guided by the bees’ exploration and exploitation mechanisms to maximize the quality of the honey within the hive. In this chapter, first, the foraging behavior of real honey bees is summarized and then, the details of Artificial Bee Colony algorithm about how it mimics the foraging behavior is provided. In the third section, an application of ABC algorithm is carried out on a software engineering problem, software defect prediction.