The world is becoming more prominent and more complex every day. The resources are limited and efficiently use them is one of the most requirement. Finding an Efficient and optimal solution in complex problems needs to practical methods. During the last decades, several optimization approaches have been presented that they can apply to different optimization problems, and they can achieve different performance on various problems. Different parameters can have a significant effect on the results, such as the type of search spaces. Between the main categories of optimization methods (deterministic and stochastic methods), stochastic optimization methods work more efficient on big complex problems than deterministic methods. But in highly complex problems, stochastic optimization methods also have some issues, such as execution time, convergence to local optimum, incompatible with distributed systems, and dependence on the type of search spaces.

Therefore this thesis presents a distributed and lightweight metaheuristic optimization method (MICGA) for complex problems focusing on four main tracks. 1) The primary goal is to improve the execution time by MICGA. 2) The proposed method increases the stability and reliability of the results by using the multi-population strategy in the second track. 3) MICGA is compatible with distributed systems. 4) Finally, MICGA is applied to the different type of optimization problems with other kinds of search spaces (continuous, discrete and order based optimization problems).

MICGA has been compared with other efficient optimization approaches. The results show the proposed work has been achieved enough improvement on the main issues of the stochastic methods that are mentioned before.