الفهرس 
Chapter 1: IntroductionOnly 14 pages are availabe for public view
 |  | from | 116 |  |  |
| | | from | 116 | | |
Abstract
During the past decades, numerous modern control methods such as fuzzy control, adaptive control have been introduced. Despite of that, PID controller is still the most widely used controller in the industry because of its simplicity and robustness. Finding the optimal parameters of PID controller is not a straightforward problem especially in non-linear control system as in liquid level control systems. So, several methods have been proposed to tune PID controller. Recently, many soft computing evolutionary algorithms such as differential evolution, genetic algorithm and particle swarm optimization have been employed to tune PID controller in various plants. However, each algorithm has its shortcoming. Tuning of PID controller as an optimization problem can be classified into two techniques: single objective and multiobjective (MO). Designing of MO-PID controller is considered a recent trend when designing PID controller because it gives a set of solutions with different trade-offs among different problem objectives. Decision Making (DM) is necessary for selecting the best final solution. To be able to carry out DM effectively, graphical presentation, can be helpful tool in the analysis. selection the appropriate multiobjective algorithm that can get diversity in the solutions is considered an important step in the design. Also, selection of visualization tool that help the DM can affected significantly on the analyzing the results. The main contribution of this thesis is introducing a new algorithm used to find the optimal parameter of PID controller. The proposed algorithm is based on hybridization of two soft computing evolutionary algorithms. The proposed PID controller is used to control three tanks liquid level system which is complex control system. A comparative study with 5 algorithms is analyzed. The results show that the proposed algorithm can tune PID controller efficiently. The result of the proposed controller has least standard deviation and best convergence characteristic algorithm. To validate the proposed algorithm, it is tested on 12 benchmark functions. A comparison with 6 algorithms is reported. These results show that proposed algorithm outperform the other algorithms in many cases. Also, it has acceptable convergence characteristics. A design of multiobjective PID controller using non-dominated sorting genetic algorithm II (NSGA-II) is introduced. A comparative study with 4 algorithms is analyzed .The results show that NSGA-II only could get various solutions. For selection a solution, a visualization tool called Level Diagram (LD) is used. The results show that this tool improves the process of DM.