Genetic Heritage Evolution by Stochastic Transmission in the optimal design of water distribution networks

Authors:
Andrea Bolognesi;Cristiana Bragalli;Angela Marchi;Sandro Artina
Affiliations:
DISTART, University of Bologna, Viale Risorgimento 2, 40136 Bologna, Italy;DISTART, University of Bologna, Viale Risorgimento 2, 40136 Bologna, Italy;DISTART, University of Bologna, Viale Risorgimento 2, 40136 Bologna, Italy;DISTART, University of Bologna, Viale Risorgimento 2, 40136 Bologna, Italy
Venue:
Advances in Engineering Software
Year:
2010

Citing 3
Cited 1

A gradient algorithm for the analysis of pipe networks

Computer applications in water supply: vol. 1---systems analysis and simulation
Genetic algorithms + data structures = evolution programs (2nd, extended ed.)

Genetic algorithms + data structures = evolution programs (2nd, extended ed.)
Effective Relaxations and Partitioning Schemes for Solving Water Distribution Network Design Problems to Global Optimality

Journal of Global Optimization

Network infrastructure design with a multilevel algorithm

Expert Systems with Applications: An International Journal

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper proposes a new model named GHEST, a multi-population evolutionary-strategy-like algorithm applied to the optimal design of water distribution networks (WDN). GHEST hunts for the optimal solution by means of two different complementary processes. The first one, synthesizes and transmits the genetic patrimony (heritage) of the parent solutions using their statistical indicators. The second one, called ''shuffle'', avoids the search to get stuck in local minima whenever the evolutionary potential of the population appears to be exhausted. GHEST makes use of hydraulic network solver EPANET 2. Tests carried out on classical WDN optimal design problems are shown for three small and well-known networks and for a large-size one. Performances exhibited, in terms of minimum cost, are equal or better than those found in previous works (where directly comparable). The algorithm has been tested with different setups, achieving good results for almost all of them. Its performance can be particularly appreciated in large-size optimization problems as evidenced by results on Balerma network, where a new minimum cost has been set and the evaluation number to reach the former minimum has been decreased by about 35 times. Results are supported by an extensive comparison with previous works on the benchmark networks here tested.