O.R. Applications Tabu search algorithms for water network optimization

In this paper we propose a tabu search algorithm to find the least-cost design of looped water distribution networks. The mathematical nature of this optimization problem, a nonlinear mixed integer problem, is at the origin of a multitude of contributions to the literature in the last 25 years. In fact, exact optimization methods have not been found for this type of problem, and, in the past, classical optimization methods, like linear and nonlinear programming, were tried at the cost of drastic simplifications. Tabu search is a valuable heuristic technique for solving problems cast in combinatorial form. This is based on the human memory process and uses an iterative neighborhood search procedure in an attempt to avoid becoming trapped in local optima. The use of such a heuristic procedure to solve the aforementioned problem needs particular tailoring to produce high quality solutions. In this paper we present the essential features of the algorithm and the results obtained when it is applied to some of the classical water distribution network case studies appearing in the literature. The results are very promising and demonstrate the usefulness of tabu search algorithms in solving this kind of optimization problem. � 2003 Elsevier B.V. All rights reserved.

[1]  D. B. Khang,et al.  Correction to “A two‐phase decomposition method for optimal design of looped water distribution networks” by Okitsugu Fujiwara and Do Ba Khang , 1991 .

[2]  U. Shamir,et al.  Design of optimal water distribution systems , 1977 .

[3]  D. B. Khang,et al.  A two‐phase decomposition method for optimal design of looped water distribution networks , 1990 .

[4]  Thomas M. Walski,et al.  Water Distribution Systems: Simulation and Sizing , 1990 .

[5]  Ulrich Faigle,et al.  Some Convergence Results for Probabilistic Tabu Search , 1992, INFORMS J. Comput..

[6]  Dragan Savic,et al.  Improved design of “Anytown” distribution network using structured messy genetic algorithms , 1999 .

[7]  Graeme C. Dandy,et al.  Genetic algorithms compared to other techniques for pipe optimization , 1994 .

[8]  Kaj Madsen,et al.  Optimization of pipe networks , 1991, Math. Program..

[9]  S. Narasimhan,et al.  Optimal design of water distribution system using an NLP method , 1997 .

[10]  E. Downey Brill,et al.  Optimization of Looped Water Distribution Systems , 1981 .

[11]  G. Loganathan,et al.  Design Heuristic for Globally Minimum Cost Water-Distribution Systems , 1995 .

[12]  David E. Goldberg,et al.  Genetic Algorithms in Pipeline Optimization , 1987 .

[13]  Dragan Savic,et al.  Genetic Algorithms for Least-Cost Design of Water Distribution Networks , 1997 .

[14]  O. Fujiwara,et al.  A modified linear programming gradient method for optimal design of looped water distribution networks , 1987 .

[15]  Ian C. Goulter,et al.  Optimal urban water distribution design , 1985 .

[16]  Maria da Conceição Cunha,et al.  Hydraulic infrastructures design using simulated annealing , 2001 .

[17]  A. Ben-Tal,et al.  Optimal design of water distribution networks , 1994 .

[18]  Thomas M. Walski,et al.  State-of-the-Art Pipe Network Optimization , 1985 .

[19]  I. Goulter,et al.  Implications of Head Loss Path Choice in the Optimization of Water Distribution Networks , 1986 .

[20]  M. Hanif Chaudhry,et al.  Closed-Conduit Flow , 1981 .

[21]  A. Simpson,et al.  An Improved Genetic Algorithm for Pipe Network Optimization , 1996 .

[22]  Johannes Gessler.  Pipe Network Optimization by Enumeration , 1985 .

[23]  Maria da Conceição Cunha,et al.  Water Distribution Network Design Optimization: Simulated Annealing Approach , 1999 .

[24]  U. Shamir,et al.  Analysis of the linear programming gradient method for optimal design of water supply networks , 1989 .

[25]  P. Bhave,et al.  GLOBAL OPTIMUM TREE SOLUTION FOR SINGLE-SOURCE LOOPED WATER DISTRIBUTION NETWORKS SUBJECTED TO A SINGLE LOADING PATTERN , 1993 .