Loading…
GALAXY: A new hybrid MOEA for the optimal design of Water Distribution Systems
A new hybrid optimizer, called genetically adaptive leaping algorithm for approximation and diversity (GALAXY), is proposed for dealing with the discrete, combinatorial, multiobjective design of Water Distribution Systems (WDSs), which is NP‐hard and computationally intensive. The merit of GALAXY is...
Saved in:
Published in: | Water resources research 2017-03, Vol.53 (3), p.1997-2015 |
---|---|
Main Authors: | , , |
Format: | Article |
Language: | English |
Subjects: | |
Online Access: | Get full text |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Summary: | A new hybrid optimizer, called genetically adaptive leaping algorithm for approximation and diversity (GALAXY), is proposed for dealing with the discrete, combinatorial, multiobjective design of Water Distribution Systems (WDSs), which is NP‐hard and computationally intensive. The merit of GALAXY is its ability to alleviate to a great extent the parameterization issue and the high computational overhead. It follows the generational framework of Multiobjective Evolutionary Algorithms (MOEAs) and includes six search operators and several important strategies. These operators are selected based on their leaping ability in the objective space from the global and local search perspectives. These strategies steer the optimization and balance the exploration and exploitation aspects simultaneously. A highlighted feature of GALAXY lies in the fact that it eliminates majority of parameters, thus being robust and easy‐to‐use. The comparative studies between GALAXY and three representative MOEAs on five benchmark WDS design problems confirm its competitiveness. GALAXY can identify better converged and distributed boundary solutions efficiently and consistently, indicating a much more balanced capability between the global and local search. Moreover, its advantages over other MOEAs become more substantial as the complexity of the design problem increases.
Key Points
A new and efficient hybrid algorithm is proposed for solving the discrete, multiobjective design of Water Distribution Systems
The method is robust and easy‐to‐use by alleviating the parameterization issue of Multiobjective Evolutionary Algorithms to a great extent
The hybrid algorithm outperforms three baseline methods in terms of exploration and exploitation consistently based on comparative studies |
---|---|
ISSN: | 0043-1397 1944-7973 |
DOI: | 10.1002/2016WR019854 |