Evolutionary multi-objective design optimisation with real life uncertainty and constraints

This paper presents the application of a robust evolutionary multi-objective evaluation technique for design optimisation in constrained problems with real life uncertainty. The design optimisation approach handles uncertainty using statistical formulations. The approach handles uncertainties associated both with design variables and design criteria. Constraints violation within the neighbourhood of a design is considered as part of a measurement for degree of feasibility and robustness of a solution. An industrial case study is presented on long product rolling system to minimise thermal effects and mechanical stress. The case study incorporates real life design variable tolerance and model sensitivity and process uncertainties.