Studies in heuristics for the annual crop planning problem.

Abstract

Increase in the costs associated with agricultural production and the limited availability of resources have amplified the need for optimized solutions to the problem of crop planning. The increased costs have imparted negatively on both the cost of production as well as the sale prices of finished products to consumers, with the resultant effects on the socio-economic livelihoods of people around the world. This has increased the burden of poverty, malnutrition, diseases and other types of social problems. The limited availability of land, irrigated water and other resources in crop planning therefore demand optimal solutions to the problem of crop planning, in order to maintain the desired level of profitable outputs that do not strain available resources while still meeting the demands of consumers. Incidentally, the current situation is such that crop producers are required to generate more output per area of crops cultivated within the ambit of the available resources for crop production. This creates a great challenge both for farmers and researchers. Interesting, the problem is essentially an optimization problem hence a challenge to researchers in mathematical and computing science. Notably within the agricultural sector, achieving efficient use of irrigated water demands that optimized solutions be found for its usage during crop planning and production. Incidentally, increase in population growth and limited availability of fresh water has increased the demand of fresh water supply from all sectors of the economy. This has increased the pressure on the agricultural sector as being one of the primary users of fresh water supply to use irrigated water more efficiently. This is to minimize excessive water wastage. It has therefore become very important that optimized solutions be found to the allocation and use of the irrigated water, for water conservational purposes. This is also a very essential key to crop planning decisions. Therefore, in order to determine good solutions to crop planning decisions, this study dwells on a fairly new but important area of agricultural planning, namely the Annual Crop Planning (ACP) problem which essentially focuses at the level of an irrigation scheme. The study presents a model of the ACP problem that helps to determine solutions to resource allocations amongst the various competing crops that are required to be grown at an irrigation scheme within a year. Both new and existing irrigation schemes are considered. Determining solutions for an ACP problem requires that the requirements and constraints presented by crop characteristics, climatic conditions, market demand conditions and the variable costs associated with agricultural production are observed. The objective is to maximize the total gross profits that can be earned in producing the various crops within a production year. Due to the complexity involved in determining solutions for an ACP problem, exact methods are not researched in this study. Rather, to determine near-optimal solutions for this -Hard optimization problem, this research introduces three new Local Search (LS) metaheuristic algorithms. These algorithms are called the Best Performance Algorithm (BPA), the Iterative Best Performance Algorithm (IBPA) and the Largest Absolute Difference Algorithm (LADA). The motivation for implementing these algorithms is to investigate techniques that can be used to determine effective solutions to difficult optimization problems at low computational costs. This study also investigates the performances of three recently introduced swarm intelligence (SI) metaheuristic algorithms in determining solutions to the ACP problems studies. These algorithms have shown great strength in providing competitive solutions to similar optimization problems in literature, hence their use in this work. To the best of the researchers’ knowledge, this is the first work that reports comparative study of the performances of these particular SI algorithms in determining solutions to a crop planning problem. Interesting results obtained and reported herein show the viability, effectiveness and efficiency of incorporation proven metaheuristic techniques into any decision support system that will help determine solutions to the ACP problem.