Many chemical processes are modeled using stochastic simulations (MC, KMC, SSA, etc). The optimization of these processes is very challenging due to: (a) The model is in a form of simulation, (b) The computational cost of the simulation is high and (c) There is a statistical noise associated with the simulations. The objective of this work is to accelerate the optimization by using a very crude model in most part of the optimization procedure. Evolutionary algorithm (EA) and artificial chemical process (ACP) algorithm [1] are modified using concepts of optimal comparison and stochastic ruler during the search. The modified algorithms accelerate the optimization tremendously. Dynamic and hybrid dynamics optimization of a complex nano-particle formation is solved using this methodology ([2], [3]). The algorithms are compared in terms of robustness and speed in finding the global optimum as a function of the model noise (from accurate simulations to very crude simulations with very large noise).

[1] R. Irizarry, LARES: An Artificial Chemical Process Approach for Optimization (2004) Evolutionary Computation Journal, 12 (4), 435-460.

[2] R. Irizarry, Fast Monte Carlo Methodology for Multivariate Particulate Systems-I: Point Ensemble Monte Carlo (2008) Chemical Engineering Science 63, 95-110.

[3] R. Irizarry, Hybrid Dynamic Optimization using Artificial Chemical Process: Extended LARES-PR (2006) Industrial & Engineering Chemistry Research, 45, 8400-8412.