Comparison of Swarm Optimization and Genetic Algorithm for Mobile Robot Navigation 47 Comparison of Swarm Optimization and Genetic Algorithm for Mobile Robot Navigation

Swarm optimization, swarm intelligence and swarm robotics are the fields considering a group of relatively simple individuals able cooperate to perform complex tasks, in decentralized manner. The inspiration is found in the first line within animal societies, such as birds, ants and bees. Social insects exhibit successful behavior in performing complex tasks on the level of the group, and are able to eliminate noise, errors, failure of swarm members. These swarms are robust, able to adapt to constant environmental changes in conditions of limited communications among members and lack of global data. In the context of swarm optimization, the example of Dorigo’s “Ant Colony Optimization “ (ACO) and Kennedy ad Eberhart “Particle swarm Optimization” (PSO) are most known examples of applying swarm-based concepts to development of optimization algorithms able to cope with hard optimization problems. These algorithms are justifiably called swarm algorithms, because they are run asynchronously and in decentralized manner (Benni, 2004). They also mimic the stigmergic (communication by dynamically changing environment) behavior of swarm of insects. PSO is inspired by flocking behavior of the birds searching for food. Although PSO shares many common attributes with the field of Genetic Algorithms (GA), such as stochastic nature, population of solution candidates, PSO methods, unlike GA use a kind of cooperation between the particles to drive the search process. PSO methods have no evolutionary operators like crossover and mutation. Each particle keeps track of its own best solution, and the best solution found so far by the swarm. It means that the particles posses own and collective memory, and are able to communicate. The difference between the global best and personal best is used to direct particles in the search space. ACO employs the search process that is inspired by the collective behavior of trail deposit and follow-up, which is observed within real ant colonies. A colony of simple agents, the ants, communicates indirectly via dynamic modifications of their environment (trails of pheromones) and thus proposes solution to a problem, based their collective experience. Honey Bees Mating Algorithm (HBMA) can also be observed as a typical swarm based approach to optimization. The algorithm is inspired by behavior of eusocial insects, which are characterized by three main features: cooperation among adults in brood care and nest construction, overlapping of at least two generations, and reproductive division of labour, 3