The analysis of complex systems during the design process often requires a decomposition of the problem into coupled subsystems which must be solved iteratively. Because this analysis is likely to be performed numerous times within an optimization framework, sequencing the subsystem modules to reduce the time and cost required to converge the analysis becomes important. This sequencing takes the form of an optimization problem which has currently only been suboptimally solved by heuristic methods. The objective of this optimization is the minimization of elements which contribute to the amount of iteration required to converge the solution. In an effort to optimize the module sequence, a method is developed to solve the sequencing problem using genetic algorithms. The method is partially successful with small systems of less than 10 modules, and some validation was found for the application of the method, but it becomes ineffective with larger systems because the percentage of possible strings not representing a valid sequence increases rapidly with the number of modules. The nature of the sequencing problem, results of trial cases, and future work are discussed.
[1]
Prabhat Hajela,et al.
Genetic search - An approach to the nonconvex optimization problem
,
1989
.
[2]
J. Sobieszczanski-Sobieski,et al.
A linear decomposition method for large optimization problems. Blueprint for development
,
1982
.
[3]
Christina Bloebaum,et al.
Heuristic decomposition for non-hierarchic systems
,
1991
.
[4]
Christina Bloebaum,et al.
NON-HIERARCHIC SYSTEM DECOMPOSITION IN STRUCTURAL OPTIMIZATION
,
1992
.
[5]
C. N. Potts,et al.
An algorithm for the single machine sequencing problem with precedence constraints
,
1980
.
[6]
David E. Goldberg,et al.
Genetic Algorithms in Search Optimization and Machine Learning
,
1988
.
[7]
James L. Rogers.
A knowledge-based tool for multilevel decomposition of a complex design problem
,
1989
.