Exploring the Bee Hive Metaphor as a Model for Problem Solving: Search, Optimisation, and More

Abstract

One of the great challenges attracting researchers of various disciplines has been the question: how to solve it? Problem solving has been an important topic of research e.g. in artificial intelligence and many remarkable results have been achieved there. However, the challenge is too complex that only one discipline, or one paradigm, or one approach can prevail. Within artificial intelligence, there has been a natural tendency to relate somehow artificial problem solvers to human ones: either in the way they behave, or in the way they actually do it. However, nature seems to offer a much broader wealth of inspiration. Among other sources of inspiration, social insects seem to play a distinct role. Indeed, their behaviour is interesting not only individually, but also in a collective, or a colony, or a swarm. From among the kinds of social insects, ants have gained perhaps the biggest attraction of problem solving researchers. Their ability to solve classes of optimisation tasks with a help of pheromones has become infamous. Only relatively recently, other kinds of social insects began to draw attention of researchers. We shall devote our attention to honey bees (apis mellifera). Honey bees are relative newcomers, although the very idea of taking inspiration from a bee hive model to represent knowledge for a knowledge based system can be traced back to 1986 (Bullock, 1986). We are interested in taking inspiration from honey bees for devising of approaches to problem solving and in particular to automated, or computer based problem solving. Behaviour of honey bees is a subject of study of other disciplines, in particular of biologists. Their studies have proven to be extremely useful (Beekman, 2007), (Biesmeijer, 2001), (Bonabeau, 1996), (Camazine, 1991), (Selley, 1991), (Zhang, 2006). Without them, we would not have the underlying knowledge on how honey bees behave in nature. From the problem solving perspective, much of the research has been concentrated on the optimisation task, perhaps in reference to the success of ant colonies (Karaboga, 2005), (Pham, 2006), (Teodorovic, 2005), (Tovey, 2005), (Wong, 2008).