Water Network Design by MINLP

Cristiana Bragalli(cristiana.bragallimail.ing.unibo.it)
Claudia D'Ambrosio(c.dambrosiounibo.it)
Jon Lee(jonleeus.ibm.com)
Andrea Lodi(andrea.lodiunibo.it)
Paolo Toth(paolo.tothunibo.it)

Abstract: We propose a solution method for a water-network optimization problem using a nonconvex continuous NLP (nonlinear programming) relaxation and a MINLP (mixed integer nonlinear programming) search. Our approach employs a relatively simple and accurate model that pays some attention to the requirements of the solvers that we employ. Our view is that in doing so, with the goal of calculating only good feasible solutions, complicated algorithmics can be confined to the MINLP solver. We report successful computational experience using available open-source MINLP software on problems from the literature and on difficult real-world instances.

Keywords: mixed integer nonlinear programming; modeling; open-source software; real-world instances

Category 1: Integer Programming ((Mixed) Integer Nonlinear Programming )

Category 2: Applications -- Science and Engineering (Civil and Environmental Engineering )

Citation: IBM Research Report RC24495, 02/2008

Download: [PDF]

Entry Submitted: 03/07/2008
Entry Accepted: 03/07/2008
Entry Last Modified: 03/07/2008

Modify/Update this entry