A new heuristic approach for the design of water distribution networks invo
lving a robust fuzzy linear program optimization in which the capital costs
of the network are minimized while maintaining the nodal heads at demand n
odes within a satisfactory region as defined by the customers at those node
s is presented. Iterative interaction between the fuzzy linear program and
a network solver is used to ensure hydraulic consistency. Level of service
is modelled by the residual nodal head available at demand nodes with the s
ubjective nature of customers' satisfaction with the nodal head being repre
sented through fuzzy sets which reflect more realistically consumers' attit
udes toward pressure variations in the supply of water. Non-probabilistic u
ncertainty in the demand is modelled by a trapezoidal possibility distribut
ion function. The model is demonstrated by application to an example networ
k.