Studies in Swarm Intelligence techniques for annual crop planning problem in a new irrigation scheme

Abstract

Annual crop planning (ACP) is an NP-Hard type optimisation problem in agricultural planning. It involves finding optimal solutions for the seasonal allocations of a limited amount of agricultural land among the various competing crops that need to be grown on it. This study investigates the effectiveness of employing three relatively new Swarm Intelligence (SI) techniques in determining solutions to an ACP problem at a new irrigation scheme. The SI metaheuristics studied include Cuckoo Search (CS), Firefly Algorithm (FA), and Glow-worm Swarm Optimisation (GSO). The solutions determined by these SI techniques are compared against the solutions of Genetic Algorithm (GA), another population-based metaheuristic technique. This helps to determine the relative merits of the solutions found by the SI techniques. The results show that the SI algorithms delivered solutions superior to those of GA in determining solutions to the ACP problem at a new irrigation scheme.

Translated abstract

OPSOMMING Jaarlikse oesbeplanning is 'n NP-Hard soort optimiseringsprobleem in landbou beplanning. Dit behels die bepaal van optimale oplossing vir die seisoenale toekenning van 'n beperkte hoeveelheid landbougrond aan die verskeie mededingende gewasse. Hierdie artikel ondersoek die doeltreffendheid van drie relatiewe nuwe Swerm Intelligensie tegnieke om oplossings tot oesbeplanning by 'n nuwe besproeiingskema te vind. Die Swem Intelligensie tegnieke wat ondersoek is, is die Koekoek Soekmetode, die Vuurvliegie Algoritme en die Gloei-wurm Swerm Optimisering tegnieke. Die oplossings deur hierdie tegnieke verkry is vergelyk met dié verkry met die tradisionele Genetiese Algoritme. Dié vergelyking help om die relatiewe voordele van die nuwe Swerm Intelligensie tegnieke te bepaal. Die resultate toon dat die voorgestelde tegnieke beter oplossings as die tradisionele Genetiese Algoritme benadering gelewer het.