论文信息 - Directed Multiple Objective search of design spaces using Genetic Algorithms and neural networks

Directed Multiple Objective search of design spaces using Genetic Algorithms and neural networks

When performing search using Multiple Objective Genetic Algorithms (MOGAs) the aim is to maintain a diverse range of solutions whilst trying to converge these solutions onto the trade-off surface. In this paper a method to focus a MOGA search onto specific areas of this trade-off surface is investigated. Using interaction with a designer or decision maker his general preferences can be captured during search and modelled using an Artificial Neural Network (ANN). This allows the designer to direct the search of the design space into the regions of most benefit.

Pratyush Sen | David S. Todd | P. Sen | D. Todd

[1] Donald E. Grierson. Conceptual Design Using Emergent Computing Techniques , 1996 .

[2] Geoffrey E. Hinton,et al. Learning representations by back-propagating errors , 1986, Nature.

[3] Jian-Bo Yang,et al. Multiple Criteria Decision Support in Engineering Design , 1998 .

[4] Pekka J. Korhonen,et al. Multiple criteria decision support: The state of research and future directions , 1992, Comput. Oper. Res..

[5] Pratyush Sen,et al. A Multiple Criteria Genetic Algorithm for Containership Loading , 1997, ICGA.

[6] P. Werbos,et al. Beyond Regression : "New Tools for Prediction and Analysis in the Behavioral Sciences , 1974 .

[7] Larry Bull,et al. Model-based evolutionary computing: a neural network and genetic algorithm architecture , 1997, Proceedings of 1997 IEEE International Conference on Evolutionary Computation (ICEC '97).