| Autor: |
Kirstin Day, Richard J. Balling, Erik J. Schaumann |
| Rok vydání: |
1998 |
| Předmět: |
|
| Zdroj: |
7th AIAA/USAF/NASA/ISSMO Symposium on Multidisciplinary Analysis and Optimization. |
| DOI: |
10.2514/6.1998-4974 |
| Popis: |
Most optimization strategies require that there be only one objective function. When multiple objectives exist for a design problem, the objectives must be combined into a single fitness value using weighting factors for the respective objectives. It can be very difficult to decide which weighting factors should be used.. A method of optimization for problems with more than one objective function which eliminates the need to weight the objectives is presented. This method uses a pareto fitness function which assigns a single fitness value to each of the designs based on all of the objective functions. A genetic algorithm is used to find the pareto optimal set of designs. The pareto optimal set contains all the designs which present the best combination of objective values regardless of weighting factors. The pareto set of designs can be presented to the decision-makers who ultimately decide the relative importance of the objectives. Two examples of this method are summarized. The first example examines the design and construction of a multi-story concrete building. The second example determines the optimal land zoning and street types for an area experiencing tremendous growth. |
| Databáze: |
OpenAIRE |
| Externí odkaz: |
|
