Recasting nonlinear differential equations as S-systems: a canonical nonlinear form

Abstract

An enormous variety of nonlinear differential equations and functions have been recast exactly in the canonical form called an S-system. This is a system of nonlinear ordinary differential equations, each with the same structure: the change in a variable is equal to a difference of products of power-law functions. We review the development of S-systems, prove that the minimum for the range of equations that can be recast as S-systems consists of all equations composed of elementary functions and nested elementary functions of elementary functions, give a detailed example of the recasting process, and discuss the theoretical and practical implications. Among the latter is the ability to solve numerically nonlinear ordinary differential equations in their S-system form significantly faster than in their original form through utilization of a specially designed algorithm.

Polynomial representations of piecewise-linear differential equations arising from gene regulatory networks

“…We can reformulate the problem using the convenient terminology from systems biology [11] for which the number…”

Exploring aspects of cell intelligence with artificial reaction networks

“…In the domain of chemistry, they can directly represent (or be slightly modified to represent) all the common reaction types. They form the basis of S-systems (Savageau and Voit 1987) and are well characterised in biochemical simulations. The basic rate equation is described by Eq.…”

Temporal Patterns in Artificial Reaction Networks

“…Our previous paper provides a complete description, and verification of the ARNs accuracy and biological plausibility [1] There are many methods used to model biochemical reactions, some are very simple Boolean-based techniques, others complex quantum mechanical abstractions [13], here the two most relevant adopted techniques are described. The first is S-Systems; these have proven themselves accurate and provide a similar degree of system abstraction to an ANN.…”

“…The first is S-Systems; these have proven themselves accurate and provide a similar degree of system abstraction to an ANN. They comprise sets of ordinary differential equations (ODEs) that exploit a power law representation to approximate chemical flux [13]. Similarly to traditional rate law [13], each ODE is equal to the difference between two conceptually distinct functions; the first function includes all terms contributing to system influx, the second to decay.…”

“…They comprise sets of ordinary differential equations (ODEs) that exploit a power law representation to approximate chemical flux [13]. Similarly to traditional rate law [13], each ODE is equal to the difference between two conceptually distinct functions; the first function includes all terms contributing to system influx, the second to decay. S-systems provide simple but accurate representations of temporal dynamics, including both steady and transient state.…”