Article ID: | iaor1999874 |
Country: | Netherlands |
Volume: | 29 |
Issue: | 1/4 |
Start Page Number: | 151 |
End Page Number: | 176 |
Publication Date: | Sep 1997 |
Journal: | Engineering Optimization |
Authors: | Parmee I.C., Vekeria H., Bilchev G. |
Keywords: | search, stochastic processes, design, adaptive processes |
The paper introduces various strategies which incorporate evolutionary and adaptive search techniques. These strategies incorporate genetic algorithms (GA) and ant colony models combined within co-operating frameworks that provide a capability for decision support and optimization during whole system design and constraint satisfaction/constrained optimization during the engineering design process. The objective during whole system design is to determine an optimum initial configuration for large engineering systems. Strategies for the efficient integration of evolutionary techniques with detailed design are also introduced. Each of these areas presents specific problems to the evolutionary/adaptive search processes and the overall objective here is to identify the main areas of difficulty and provide solutions that will lead to successful integration. The paper illustrates the flexibility and utility of the various techniques when applied across the various stages of the design process, i.e. from providing decision support during the high-risk stages of preliminary design to the identification of definitive optimal solutions during the more deterministic stages of detailed design.