Abstract
A regional water supply system design problem that determines pipe and pump design parameters and water flows over a multi-year planning horizon is considered. A non-convex nonlinear model is formulated and solved by a branch-and-reduce global optimization approach. The lower bounding problem is constructed via a three-pronged effort that involves transforming the space of certain decision variables, polyhedral outer approximations, and the Reformulation Linearization Technique (RLT). Range reduction techniques are employed systematically to speed up convergence. Computational results demonstrate the efficiency of the proposed algorithm; in particular, the critical role range reduction techniques could play in RLT based branch-and-bound methods. Results also indicate using reclaimed water not only saves freshwater sources but is also a cost-effective non-potable water source in arid regions. Supplemental data for this article can be accessed at http://dx.doi.org/10.1080/0305215X.2015.1016508.
Original language | English (US) |
---|---|
Pages (from-to) | 454-475 |
Number of pages | 22 |
Journal | Engineering Optimization |
Volume | 48 |
Issue number | 3 |
DOIs | |
State | Published - Mar 3 2016 |
Keywords
- branch-and-reduce
- global optimization
- range reduction
- reformulation linearization technique
- regional water supply system
ASJC Scopus subject areas
- Computer Science Applications
- Control and Optimization
- Management Science and Operations Research
- Industrial and Manufacturing Engineering
- Applied Mathematics