Value of information in production processes : the case of aquaculture

Abstract

Production processes involve considerable uncertainty. In decision making, the availability of information concerning the random variables is important. In this study information about random variables is evaluated for an economic unit with different attitudes toward risk. Four optimization models are defined. The models differ from one another in the level of information that is assumed to exist at the time of decision. A learning mechanism about the random variables is developed and optimized over the economic unit's utility function. The value of information about random variables in the system is derived from each model. A comparison among models is then offered. The value of information is derived under alternative model specifications. The culture of Penaus vienammi, in a semi-intensive system in the central coast of Texas, is the case study. Shrimp aquaculture in Texas is new, little research has been conducted and little experience has been gained in commercial aquaculture. The uncertainty associated with raising shrimp in ponds is considerable. Shrimp production is constrained by weather conditions and the resources available to farmers. The domestic shrimp market has existed for years. All the adjoined processes such as the marketing system do work (cold houses, storage, etc.). Thus, the primary focus is on production uncertainty. The management of the production process faces classical problems of how to use the control variables in order to optimize the objective function. Using optimal control framework, the state variable vector (x) consists of two components, weight of an individual shrimp (w) and quantity (number) of shrimp per unit of pond (q). The objective of the model is a maximization of expected utility of profit above fixed costs, within a year. The output of the model is the optimal schedule for stocking and harvesting days, stocking levels and feeding rates. The problem is then solved numerically. The case study demonstrates that a monetary value can be assigned to different levels of sample information; that more risk averse decision makers (may) pay more for the information on their production process; and the number-size trade off in shrimp production is not trivial and has important implications for aquaculture profitability. (Abstract shortened with permission of author.)