Adaptive optimisation of noisy black-box functions inherent in microscopic models

For systems where exact constitutive relations are unknown, a microscopic level description can be alternatively used. As microscopic imulations are computationally expensive, there is a need for the development of robust algorithms in order to efficiently optimise such systems aking into consideration the inherent noise associated with the microscopic description. Three optimisation strategies are proposed and tested sing a stochastic reaction system as a case study. The first method generates optimal difference intervals to formulate and solve a non-linear rogram (NLP), whereas the other methods build response surface models and optimise using either a direct search algorithm changing to a teepest descent method once the optimum region is located, or sequential quadratic programming (SQP). The performance of these methods is ompared to that of a steepest descent optimisation method commonly used for response surfaces. Their effectiveness is evaluated in terms of the umber of microscale function calls and computational time. 2006 Elsevier Ltd. All rights reserved.