Output details

Stochastic sampling design using a multi-objective genetic algorithm and adaptive neural networks

Determining best sensor locations in pipe networks is a major issue for the UK water industry which is currently spending millions of pounds on installing new devices. The methodology presented here is a paradigm shift in thinking which acknowledges explicitly the existence of a number of uncertainties in the process. The paper also features the new, generic MOGA-ANN optimisation method which has been later on used to solve a range of other important problems (pump scheduling, distribution system rehabilitation and urban drainage network redesign/operation) by this and other authors.