SUPERVISORY CONTROL MENGGUNAKAN ALGORITMA GENETIK

• Main Author: Nadhir, Ahmad
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/5361
• Institution: Institut Teknologi Bandung

<p>Absrtact:<p align=\"justify\"> <br /> <br /> <br /> Genetic algorithm is one of some methods oftenly utilised to find optimum solutions of problems. The use of fitness function will determine selection of best solution, in one generation among many solution candidates. There are several parameters used in genetic algorithm, among them is crossover probability parameter Pp and mutation probability Pm. The best selection of genetic algorithm parameter can increase the performance that leads to an optimum results.<p align=\"justify\"> <br /> <br /> <br /> <br /> In designing the Proportional-Derivative controller (PD controller), controller parameters IC and Kd will be determined the appropriate. The genetic algorithm can be included in designing Proportional Derivative controller that functions in determining controller the parameter I ands which are optimum at any sampling time. The fitness value a number of candidates of controller parameter will be evaluate by genetic algorithm to be selected the best control parameter. Genetic algorithm operators (selection, cross over, and mutation) are used to generate new controller parameter candidate population. Selection of genetic algorithm parameter and fitness function parameter Pm, Pc, a,,, a2, a3 can appropriated with the problem to produce control action as expected. Supervisory control works to determine optimum parameter of genetic algorithm and fitness function parameter uses as controller.<p align=\"justify\"> <br /> <br /> <br /> <br /> The use of genetic algorithm on plant inverted wedge can increase performance on control action if compared with control without genetic algorithm. Increase of performance is shown by the decrease of settling time which is needed to achieved steady state. Optimum parameters of genetic algorithm and fitness function used in controller are Pm 0.05, Pe=0.9, a1=600, a2=10, a3=400. Optimum parameters of genetic algorithm used in supervisory control are Pm•0.01,P\'0.5.