Production Planning in Fisheries Under Uncertainty. Stochastic Optimization and Scenario Generation - A Case Study

Abstract

Uncertainty in the value chain of fisheries exists and as a result, a production model taking uncertainty into account is important for such a business to be efficient. A way to better plan under uncertain conditions is through the use of stochastic programming. The literature for which this is applied to fisheries is limited. Constructing, testing, and evaluating such a model's applications for a fishery processing plant producing dried and salted, and fresh fish is the purpose of the thesis. This is done through a case study where the method of stochastic optimization is applied. Initially, scenarios representing the underlying distributions are generated through time series models for the variables and parameters exhibiting uncertain behaviour. These scenarios are used as input values for the mathematical program representing the value chain of the fishery. The results yielded in this thesis indicate that it is indeed an increase in efficiency to be had applying such a model, although the low value of the stochastic solution (VSS) estimates makes it difficult to conclude with certainty. Consequently, it is suggested to increase the complexity of the model to better represent the whole value chain in greater detail which is expected to increase the VSS. Furthermore should different scenario generating methods be evaluated for both harvest and price to compare the stability of the results as per now they are suspected to be to somewhat unstable as indicated by their dispersion and central tendency results.