Parameter tuning of a stochastic biological simulator by metaheuristics

Abstract

In this paper we address the problem of tuning parameters of a biological model, in particular a simulator of stochastic processes. The task is defined as an optimisation problem over the parameter space in which the objective function to be minimised is the distance between the output of the simulator and a target one. We tackle the problem with a metaheuristic algorithm for continuous variables, Particle swarm optimisation, and show the effectiveness of the method in a prominent case-study, namely the mitogen-activated protein kinase cascade. © Springer-Verlag 2009.