Using fuzzy logic models to reveal farmers motives to integrate livestock, fish, and crops.
MetadataShow full item record
Bosma, R. H. 2007. Using fuzzy logic models to reveal farmers motives to integrate livestock, fish, and crops. Phd Thesis, Wageningen University.
Permanent link to cite or share this item: https://hdl.handle.net/10568/79732
External link to download this item: http://www.narcis.nl/publication/RecordID/oai%3Alibrary.wur.nl%3Awurpubs%2F359642
Rural extension services have changed paradigm and shifted to more participatory approaches, whereas in common mathematical models of farming systems, farmers’ motivation is solely represented by ‘utility maximisation’. While globally, farmers specialise, in Vietnam the rice-based systems have diversified into more sustainable integrated agriculture–aquaculture. We gathered data from 144 farms in six villages in two ecological zones of the Mekong Delta, Vietnam. Using the livelihood framework we conceptualised farmers’ decision-making in a fuzzy logic model that can deal with subjective linguistic statements through ‘if–then’ rules. The desire to improve livelihoods and diet, mainly for their children’ well-being was the farmers’ main motive for diversification. Livestock, including fish, was essential in the expansion and accumulation stages of the nuclear families’ life-course having five stages. In 10 recursive steps we developed a model of farmers’ decision-making in a transparent hierarchical tree composed of several Mamdani-based inference systems, each with its rule base. Model conceptualisation, variables selection, model structuring, and definition of linguistic values, membership functions and rule base were based on a first set of data that was completed before calibration. In a pilot, the simulation of the frequency distribution of four fish-production systems was good, but classification of individual farmers was poor. Using composed variables for land, water, labour and capital decreased the fuzziness of the inference in this pilot model. In a more elaborated three-layer model, the whole farm composition was simulated using variables for the production factors, farmers’ appreciation of prices, farmer’s know-how of 10 activities, operational variables of social motives for integration and diversification as well as for risk-taking behaviour and for rice food security. Model’s classification of individual farmers in the delta was good for the land-based activities but poor for the livestock activities. A test on the hill farmers’ dataset showed that the model was context-specific. The model’s sensitivity to the social variables determining diversification and integration was of the same magnitude as its sensitivity to product’s prices and farmer’s know-how, but smaller than its sensitivity to labour, capital and land endowment. We conclude that farmers’ decision-making can be simulated using a fuzzy logic model. In the Mekong Delta farm diversification and integration are driven by labour, income, homestead area, number of young children, index of integration, household life-course, and level of education and age of the household head, in decreasing order. The choice of a component depends on the household’s assets and specific know-how, and on marketability. Farm models that do not include family-related motivations might be less reliable than generally suggested.