Scientific Objectives

The overall goal of this project is to study the computational properties of CSNs by evolving them using methods from evolutionary computation, and to re-apply this understanding in developing new ways to model and predict real CSNs. This is achieved via a set of secondary objectives as follows:

Objective 1: Develop an understanding of the computational properties of CSNs.

In the context of this project this entails:

1.1 Understand how to create artificial CSNs (in a computational environment) that implement pre-defined computations.

1.2 Understand how CSNs scale with the complexity of the computational task. How does the minimal size (i.e. the number of di_erent kinds of particles) of CSNs scale with the complexity of the computational task?

1.3 Understand whether there are several stucturally di_erent CSNs that perform the same computation.

1.4 Understand what types of computations CSNs can e_ciently deal with and what types of computations they are not suitable for.

1.5 Investigate robustness of CSNs to perturbations (failure of components, random fluctuations of particle concentrations, changing inputs)?

1.6 Investigate the dynamics of interconnected (“crosstalking”) CSNs and how crosstalk leads to mutual control of CSNs.

1.7 Develop methods to formally check the computational properties of real and artificial CSNs.

1.8 Develop a mathematical theory of the computational properties of CSNs.

1.9 Understand the dynamical properties of the interactions between the particles in CSNs.

Objective 2: Gain new theoretical perspectives on real CSNs and to create new ways to model real CSNs.

2.1 Develop methods to evolve biologically plausible models of real CSNs.

2.2 Investigate whether real CSNs are optimal and if so what they are optimised for.

2.3 Compare performance of real CSNs with functionally equivalent simulated alternatives.

2.4 Understand evolutionary constraints and trade-o_s that are likely to have led to the observed real CSN.

Objective 3: Predict unknown components of partially known CSNs.

Objective 4: Develop open-source software packages for scientific use.

4.1 A software library that can be used to evolve and simulate CSNs.

4.2 A predictor for components of real CSNs.

Author: : root -- 17.1.2006 2:05:51

This article 4807 times visited.