Fixed points of order-reversing maps in Rn>0and chemical equilibrium

The problem of computing the equilibrium state of a reversible chemical reaction network has a natural interpretation as a fixed-point problem. Several authors have used fixed-point iterations in this context, yet there are no comprehensive investigations into the convergence of the algorithm. We address this void by studying the larger problem of the existence and uniqueness of fixed points, and the convergence of fixed-point iterations, for order-reversing maps in Rn 0. By using the Thompson metric, we are able to apply fixedpoint theorems based on the Lipschitz condition and obtain upper bounds on judiciously defined approximation errors.