Optimal portion control of natural objects with application in automated cannery processing of fish

This payer addresses the problem of optimal portioning of natural objects s uch as fish, meat, and other food items. Portions need to be optimized for reasons such as wastage reduction, regulatory requirements, consumer appeal , and aesthetic considerations. The paper presents an optimal approach for portion control. This involves cutting a set of objects into pieces by taki ng into account the density distribution of each object and grouping the pi eces into package portions according to an optimality criterion. The object ive of optimization is to make each portion as close as possible to the tar get weight of the package, within a tolerance. First, an optimization model is developed for a general formulation of the problem. Next, the portionin g process is refined by incorporating realistic assumptions and constraints that are defined for the specific type of product and portioning process. The modified model is checked again for feasibility. The model refinement p rocess is continued until a feasible optimization model that can be impleme nted on-line in an industrial plant is achieved. The procedure developed he re is applied to an industrial fish canning process. In this application, t he optimization model is refined twice to arrive at a computationally tract able implementation. A numerical example that uses real data on a batch of salmon is given to illustrate the advantages of optimal portion control; sp ecifically, improvement in filling accuracy, associated benefits of meeting regulatory requirements, and increase in production efficiency. (C) 2000 E lsevier Science Ltd. All rights reserved.