Abstract

Experimental optimisation with hardware-in-the-loop is a common procedure in engineering, particularly in cases where accurate modelling is not possible. A common methodology to support experimental search is to use one of the many gradient descent methods. However, even sophisticated and proven methodologies such as Simulated Annealing (SA) can be significantly challenged in the presence of significant noise. This paper introduces a decision support methodology based upon Response Surfaces (RS), which supplements experimental management based on variable neighbourhood search, and is shown to be highly effective in directing experiments in the presence of significant signal to noise (S-N) ratio and complex combinatorial functions. The methodology is developed on a 3-dimensional surface with multiple local-minima and large basin of attraction, and high S-N ratio. Finally, the method is applied to a real-life automotive experimental application.