The Application of Particle Swarm Optimization in Estimating Potential Evapotranspiration: A Brief Review

Nor Farah Atiqah Ahmad; Sobri Harun; Haza Nuzly Abdull Hamed; Suhaila Sahat; Siti Nooraiin Mohd Razali; Masiri Kaamin

Authors

Nor Farah Atiqah Ahmad Department of Civil Engineering, Centre for Diploma Studies, Universiti Tun Hussein Onn Malaysia, Edu Hub Pagoh, Johor, Malaysia
Sobri Harun School of Civil Engineering, Faculty of Engineering, Universiti Teknologi Malaysia, Skudai, Johor, Malaysia
Haza Nuzly Abdull Hamed School of Computing, Faculty of Engineering, Universiti Teknologi Malaysia, Skudai, Johor, Malaysia
Suhaila Sahat Department of Civil Engineering, Centre for Diploma Studies, Universiti Tun Hussein Onn Malaysia, Edu Hub Pagoh, Johor, Malaysia
Siti Nooraiin Mohd Razali Department of Civil Engineering, Centre for Diploma Studies, Universiti Tun Hussein Onn Malaysia, Edu Hub Pagoh, Johor, Malaysia
Masiri Kaamin Department of Civil Engineering, Centre for Diploma Studies, Universiti Tun Hussein Onn Malaysia, Edu Hub Pagoh, Johor, Malaysia

Keywords:

Hydrological process, artificial intelligence, potential evapotranspiration, particle swarm optimization

Abstract

In hydrological cycle, evapotranspiration (ET) is one of the tedious processes to measure. This has caused a massive development of empirical estimation models and the most accurate is Food and Agricultural Organization-56 Penman-Montieth model (FPM-56). The setback of this model is its data demanding which is not applicable at data scarce region and more simple models are preferable. To that avail, evolution of optimization from soft computing, enhancing the performance of simpler empirical models in estimating ET. This paper highlights the application of particle swarm optimization (PSO) in catering the estimation for potential evapotranspiration (ETp). Although the number of papers in literature related to PSO application in hydrology or anyother areas increases exponentially, the concerns is soft computing models keep advancing gambling the validity of today’s model improvement such ET estimation empirical model. To have a model that pertinent for a long time still needs a calibration from physical direct measurement. Despite all the arguments, a comprehensive AI algorithm is yet to come.