Some modified stochastic global optimization algorithms with applications
Year: 1994
Type: dissertation
Abstract: Stochastic methods for global optimization problems with continuous variables have\nbeen studied. Modifications of three different algorithms have been proposed. These are (1)\nMultilevel Single Linkage (MSL), (2) Simulated Annealing (SA) and (3) Controlled Random\nSearch (CRS). We propose a new topographical Multilevel Single Linkage (TMSL)\nalgorithm as an extension of MSL. TMSL performs much better than MSL, especially in\nterms of number of function evaluations. A new aspiration based simulated annealing algorithm\n(ASA) has been derived which enhances the performance of SA by incorporating an\naspirat.ion criterion. We have also proposed two new CRS algorithms, the CRS4 and CRS5\nalgorithms, which improve the CRS algorithm both in terms of cpu time and the number\nof function evaluations. The usefulness of the Halton and the Hammersley quasi-random\nsequences in global optimization has been investigated. These sequences are frequently\nused in numerical integration in the field of Bayesian statistics. A useful property of the\nquasi-random sequences is that they are evenly distributed and thus explore the search\nregion more rapidly than pseudo-random numbers.\nComparison of the modified algorithms with their unmodified versions is carried out on\nstandard test problems but in addition a substantial part of the thesis consists of numerical\ninvestigations of 5 different practical global optimization problems. These problems are as\nfollows:\n(1) A nonlinear continuous stirred tank reactor problem.\n(2) A chemical reactor problem with a bifunctional catalyst.\n(3) A pig-liver likelihood function.\n(4) Application and derivation of semi-empirical many body interatomic potentials.\n(5) A optimal control problem involving a car suspension system.\nCritical comparisons of the modified and unmodified global optimization algorithms\nhave been carried out on these problems. The methods applied to these problems are\ncompared from the points of view of reliability in finding the global optimum, cpu time\nand number of function evaluations.
Author: M. Montaz Ali
Institution: Loughborough University
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W754624823
