Application of ILS-RVND for the Irrigation Problem using Center Pivot

This paper proposes a metaheuristic approach to the Center Pivot Irrigation Problem, whose daily irrigation depth decision aims to minimize operating costs by ensuring soil moisture above the critical limit throughout the crop cycle. As contributions, the following are presented: (i) four domain-specific neighborhood structures; (ii) a Lookahead construction heuristic with anticipation of future cost; (iii) the ILS+RVND combination applied to the problem; and (iv) three new instances based on real climate data from Unaí-MG for corn, cotton and sugarcane crops. The experiments demonstrate that ILS+RVND achieves the known optimal solution in all instances with a zero gap or less than 0.08%, in computational time less than 3 minutes.