Artificial atom algorithm and its application to discrete problems

In the real world, there are many areas where the ideal solution of a problem is sought. These areas include engineering, medicine, dietetic, logistics and electronics. In order to solve optimization problems in these different areas, both classical mathematical methods and heuristic methods can be preferred. On the other hand, heuristic algorithms were developed by inspired from nature are popular methods for the solution of optimization problems with their run-time speed and accuracy. Recently, a new one has been added to these methods. It is also called meta-heuristic method, which is contained in multiple intuitive techniques. The meta-heuristic method, which was recently introduced into the literature and developed by inspired the compounding processes of atoms, is the subject of this thesis. This algorithm which was developed by modeling chemical ionic bond and covalent bond processes, was applied to discrete optimization problems and tried to test the performance of the algorithm in this area. In the performance of an optimization algorithm, the success in discrete optimization problems is an important criterion. For this reason, firstly, the traveling salesman problem which has widespread use in the literature has been studied. The approach which was applied to small scale and symmetric problems, was compared with frequently-used meta-heuristic methods such as tabu search, particle swarm optimization, ant colony optimization, genetic algorithms. Then, the performance of the algorithm was monitored on an optimization problem which aims to obtain a nutrition schedule consisting of food according to the daily energy requirement of person. In the study, the daily energy requirement was calculated considering the physical, physiological and sociological characteristics of the individual. Finally, an optimization was performed to provide efficient elevator service to its users by controlling multiple elevators from a center in the building. The obtained results were given comparatively. In this thesis, successful results for the applications were achieved at reasonable times.