Many problems in engineering can be traced back to optimization, however the objective functions often have many local minima, so the usual, gradient based optimisation algorithms cannot be used. In the last few years methods, that use natural behaviour to solve these problems, were commonly used. A few of these solutions are the genetic algorithms, swarm-intelligence methods, and the cuckoo search.
This thesis shows that functions, that have many local minima can be optimised with the modified cuckoo search, implemented in MATLAB environment. The efficiency of the algorithm is tested on the commonly used test functions, and with these functions a comparison is made between the modified cuckoo search and the genetic algorithm in the matter of efficiency, runtime, and dependence of the size of the population.
The third part of this thesis is about the optimization of urban traffic control with the modified cuckoo search. Meanwhile the cycle times of the traffic lamps in the intersections are optimized, so the permeability of the road network would be as high as possible.