Multi-objective optimisation with uncertainty

This paper looks at the implications on multi objective optimisation when the objectives are subject to uncertainty. In this case the High Efficiency Deep Grinding (HEDG) process is modelled using a Fuzzy Inference Cognitive Map (FICM) with uncertainty present. Uncertainty forms an extra objective with the user having to decide which is preferable; near optimal yet highly uncertain results or certain but inefficient results. Wheel loading forms a process constraint that is modelled as an objective within the FICM. Two Algorithms are applied to the problem, NSGA-II and SPEA. The uncertainty associated with each objective is aggregated using the FICM to give an overall uncertainty value for each solution.