Systems integration using evolutionary algorithms

This paper describes an approach to the systems integration problem using multiobjective genetic algorithms. An architecture for evolutionary systems integration is presented and the component parts discussed. An example of an aircraft gas turbine engine control system design problem is shown demonstrating aspects of the proposed architecture that allow many design objectives from different disciplines to be considered in parallel. Potential closed-loop control configurations are evaluated and compared against one another within an optimization framework. As a result of this analysis, it is shown how informed decisions may be made regarding the nature of the control employed, acceptable performance margins and elements of the engine design.