The Efficiency of Setting Parameters in a Modified Shuffled Frog Leaping Algorithm Applied to Optimizing Water Distribution Networks

This paper presents a modified Shuffled Frog Leaping Algorithm (SFLA) applied to the design of water distribution networks. Generally, one of the major disadvantages of the traditional SFLA is the high number of parameters that need to be calibrated for proper operation of the algorithm. A method fo...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Water (Basel) Ročník 8; číslo 5; s. 182
Hlavní autoři: Mora-Melia, Daniel, Iglesias-Rey, Pedro, Martínez-Solano, F., Muñoz-Velasco, Pedro
Médium: Journal Article
Jazyk:angličtina
Vydáno: Basel MDPI AG 01.05.2016
Témata:
algorithms
Calibration
Frogs
Genetic algorithms
Linear programming
New York
Optimization
Simulation
statistical analysis
water distribution
Chile
New York
United States > US
ANW, USA, New York
Vietnam, Hanoi
ISSN:2073-4441, 2073-4441
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:This paper presents a modified Shuffled Frog Leaping Algorithm (SFLA) applied to the design of water distribution networks. Generally, one of the major disadvantages of the traditional SFLA is the high number of parameters that need to be calibrated for proper operation of the algorithm. A method for calibrating these parameters is presented and applied to the design of three benchmark medium-sized networks widely known in the literature (Hanoi, New York Tunnel, and GoYang). For each of the problems, over 35,000 simulations were conducted. Then, a statistical analysis was performed, and the relative importance of each of the parameters was analyzed to achieve the best possible configuration of the modified SFLA. The main conclusion from this study is that not all of the original SFL algorithm parameters are important. Thus, the fraction of frogs in the memeplex q can be eliminated, while the other parameters (number of evolutionary steps Ns, number of memeplexes m, and number of frogs n) may be set to constant values that run optimally for all medium-sized networks. Furthermore, the modified acceleration parameter C becomes the key parameter in the calibration process, vastly improving the results provided by the original SFLA.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
content type line 23
ISSN:2073-4441
2073-4441
DOI:10.3390/w8050182