An evolutionary approach for finding optimal trees in undirected networks

In many utilities, such as water and gas, materials are distributed through networks of pipes. For efficiency, many such distribution networks are constructed as trees. The cost of construction and operation of these is generally a complex function of the edges which are used, so that it is impossible to use conventional algorithms for the optimisation of trees. This paper presents a method for identifying a tree which is close to optimal. This evolutionary method is based on ideas from genetic algorithms.