OPTIMASI ALGORITMA ALGA UNTUK MENINGKATKAN LAJU KONVERGENSI

Abstract

Artificial AlgaeAlgorithm (AAA) is an optimization algorithm that has advantages of swarm algorithm model and evolution model. AAA consists of three phases of helical movement phase, reproduction, and adaptation. Helical movement is a three-dimensional movement with the direction of x, y, and z which is very influential in the rate of convergence and diversity of solutions. Helical motion optimization aims to increase the convergence rate by moving the algae to the best colony in the population. Algae Algorithm Optimization (AAA ') was tested with 25 objective functions of CEC'05 and implemented in case of pressure vessel design optimization. The results of the CEC'05 function test show that there is an increase in convergence rate at AAA ', but at worst condition of AAA' becomes less stable and trapped in local optima. The complexity analysis shows that AAA has the complexity of O (M3N2O) and AAA 'has the complexity of O (M2N2O) with M is the number of colonies, N is the number of algae individuals, and O is the maximum of the evaluation function. The results of the implementation of pressure vessel design optimization show that AAA's execution time increased 1,103 times faster than AAA. The increase in speed is due to the tournament selection process in AAA performed before the helical motion, whereas in AAA 'is done if the solution after movement is no better than before. At its best, AAA 'found a solution 4.5921 times faster than AAA. At worst, AAA 'stuck on local optima because helical movement is too focused on global best that is not necessarily global optima.