Fuzzy Decision Support System for Achieving the Target Cotton (Gossypium hirsutum L.) Yield in South-Western Punjab

Abstract

Cotton (Gossypium hirsutum L.) is one of the most important commercial crops, the productivity of which is affected by several factors. In the present study, 13 different production and management factors affecting the cotton productivity were ranked based on intensity of causing significant yield loss using Garret‟s ranking method which was integrated with fuzzy logic-based decision support system. The fuzzy data-sets were formed based on experts‟ knowledge of irrigation water quality (IWQ), soil fertility (SF) status based on soil organic carbon (SOC) content and cotton whitefly (WF) population count for the study area to achieve the target output (cotton yield). For the purpose of aggregation of fuzzy rule combinations three general types of inferences viz. Mamdani System (System type-1), Sugeno System (System type-2) and Trukamoto System (System type-3) are available. These results revealed that IWQ (mean score=77.4), SF (mean score=69.0) and cotton WF (mean score=63.8) were the major determinants of cotton productivity in south-western Punjab. Results revealed that irrigation of cotton crop with good quality irrigation water in medium to highly fertile soil yielded optimal seed cotton yield, if whitefly infestation in well below economic threshold level (ETL<6 adults leaf-1). For instance, if IWQ is marginal (Category- II), soil was medium fertile (SOC=4.0-7.5g kg-1) and whitefly population count was <ETL, 18.9 q ha-1 seed cotton yield could be achieved. Cotton yield predicted with change in soil fertility and whitefly population was highest when whitefly population was well below ETL. As the whitefly population count increased, seed cotton yield decreased drastically regardless of the soil fertility level. If the underground water used for irrigation to cotton is of poor quality, and whitefly population is well > ETL, the seed cotton yield was minimum.