SOUTHERN JOURNAL OF AGRICULTURAL ECONOMICS JULY, 1989 AN APPLICATION OF SAFETY-FIRST PROBABILITY LIMITS IN A DISCRETE STOCHASTIC FARM MANAGEMENT PROGRAMMING MODEL

s ia disio in m (MOTAD), could be used with rather simple developed, and data from farm-raised catfish modifications of widely used linear program-production were used to demonstrate its use. ming algorithms. Farm planning modelsOutcomes ofsequences of decisions which under uncertainty became significantly more satisfied chance constraints on ending cash tractable and increasingly accessible to the balances were traced through a specified time fm ma ement reseh c nit period. Discrete choice variables were The model presented in this paper employs specified due to the fixed nature of pond a method whereby certain multiple stage facilities. Recourse actions specified were sale problems can be specified to obtain feasible of production in excess of endogenously deter- plans for each stage while satisfying chance mined transfer levels or purchase of inputs to constraints on ending cash balances. It traces supplement needs of the next production through the outcome of a sequence of decisions, stage. Production activities cannot be changed but it is not a purely stochastic model in the during the planning period. Only yield sense that recourse actions are limited based variability was considered due to its impact on on results of a previous stage. Production ac-relative competitiveness among growth tivities cannot be changed; once a sequence i