Water distribution systems are complex networks of tanks, pipes, pumps and
valves. Distribution systems are designed to meet consumer demands while
satisfying performance criteria such as providing minimum pressures with
specified reliability and water quality. In recent years, a good deal of research
has gone into the application of formal optimization procedures to the design of
water distribution networks. The most successful optimization techniques to be
applied to this problem involve the use of genetic algorithms (GAs) which are a
class of algorithms based on artificial evolution. They simulate the mechanisms
of population genetics and natural selection to develop designs that are low-cost
and well adapted to the criteria that they must satisfy. GAs can be applied at
the detailed design, planning or concept design phase. This paper examines the
application of GAs to the concept design phase for a new water distribution
network that will supply water to farms and country towns in rural Australia. The
GA concept design approach considers the system capital costs such as the costs
of pipes, pump stations and tanks, as well as the projected lifetime pump
operating costs. The GA approach incorporates a hydraulic network solver to
ensure that all performance criteria for the system are met. Includes 6 references, tables, figures.