พันธภรณ์ ปัญญาภรณ์. A Temperature-less Simulated Annealing Algorithm for Optimization of Chemical Processes. Master's Degree(Chemical Engineering). King Mongkut's University of Technology Thonburi. : King Mongkut's University of Technology Thonburi, 2005.
A Temperature-less Simulated Annealing Algorithm for Optimization of Chemical Processes
Abstract:
Optimization is important in chemical engineering whose applications involve minimizing operating costs,
minimizing process contaminants, and minimizing energy consumptions. Many techniques have been proposed to
solve optimization problems. Metaheuristics are intelligent optimization methodologies which improve on local search
techniques and have a much better chance of locating global optima. The objective of this research is to propose new
metaheuristic by developing Simulated Annealing (SA) under a temperature-less criterion and comparing its results
with the existing one. A new metaheuristics called Statistical Search (SS) employs the statistical principles to replace
temperature terms. These optimization methods are generalized, requiring no customization, in order to solve scheduling
problem such as MS-MB SP, FSSP, SJSSP, and HENS. The effectiveness of SA, SS1, and SS2 are compared in term
of the quality of solution using the same amount of computation time.In order to achieve the best solution for all metaheuristics,
SS is very convenient because it does not require the tuning of parameters. This work proposes 2 types of SS, SS1 and SS2
which contain different acceptance criteria. Each scheduling algorithm is tested with the same set of 10-18 test problems.
The results show that SS2 can perform very well. SS method can overcome the inherent of SA weaknesses by monitoring the
statistical parameters in the simulation. In addition, SS2 gives the minimum percent relative error when compared with the
well-known solutions from the literature. Furthermore, SS2 was found to find global optimum in 50-60% of all the test problems.