Optimising intra-seasonal irrigation water allocation: Comparison between mixed integer nonlinear programming and differential evolution

  • B Grové Department of Agricultural Economics, University of the Free State, Bloemfontein, PO Box 339, Bloemfontein 9300, South Africa
  • MC Du Plessis Department of Computing Sciences, Nelson Mandela University, Port Elizabeth, South Africa
Keywords: intra-seasonal, water allocation, mixed integer nonlinear programming, differential evolution

Abstract

Optimising crop planning in conjunction with intra-seasonal water allocation necessitates the use of daily water budget calculations to determine the timing and amount of irrigation events, which complicates the solution of the problem to global optimality. The main objective of this research was to compare the intra-seasonal water allocation of a mixed integer nonlinear programming (MINLP) model with that of differential evolution (DE), to allocate a limited amount of water while considering irrigable area and the irrigation schedule that will maximise the total gross margin. Results show that both solution procedures adhere to economic theory of water allocation under limited water supply. The conclusion is that the MINLP model most likely achieved very near global optimality as the solutions of the two models were very close to each other. DE holds promise to solve more complex models involving risk and multiple crops.

Views
  • Abstract 101
  • PDF 36
Views and downloads are with effect from 11 January 2018
Published
2019-01-31
Section
Research paper