Author(s)

J.W. Davidson & I.C. Coulter

Abstract

An algorithm based on evolution principles is proposed for the design of layout of looped networks on rectilinear grids. The technique is similar in concept to genetic algorithms and evolution strategies which operate iteratively on a population of solutions to produce progressively better solutions. A key feature of the technique is a new solution coding scheme which overcomes some of the difficulties associated with the generation of infeasible solutions. This solution coding scheme uses continuous numbers on a [0-1] interval to represent the preference for links in a layout. The algorithm then finds a threshold value that selects the minimal number of links that will produce a feasible looped network. The method is demonstrated by application to an example problem. 1 Scope

Keywords