Cuckoo Search Optimization Metaheuristic Adjustment

Hard optimization problems that cannot be solved within reasonable time by standard, mathematical, deterministic methods are of great practical interest. Metaheuristics inspired by nature were recently successfully used for such problems. These metaheuristics are based on random Monte-Carlo search guided by simulation of some nature intelligence, especially evolution and swarm intelligence. One of the latest swarm intelligence algorithms is the Cuckoo Search Algorithm which has not yet been investigated thoroughly. For all such nature inspired algorithms fundamental issue is balance between use of good found solutions (exploration) and investigation of new areas of the search space in order to avoid being trapped in local minima (exploration). Specific of the Cuckoo Search Algorithm is exploitation/exploration based on Lévi flight i.e. combination of short and long steps according to Lévi distribution with infinite mean and variance. This plenary lecture concentrates on investigation of the Cuckoo Search Algorithm parameters adjustment and specifically sensitivity to Lévi distribution parameters. Key-Words: Cuckoo search algorithm, metaheuristic optimization, swarm intelligence, nature inspired