Computational chains for CACSD using Matlab containers

Computational chains which 'remember' that a computation has already been performed and its results are available for further use have been implemented using special classes of Matlab containers. A simple version management mechanism keeps track of any changes made to containers in the chain, and supports storage and retrieval of design data. Computational chains can be used as a basis for a more sophisticated version of management mechanism for maintaining a complete design-history database. Advantages of the computational chains become more apparent as the complexity of the work undertaken with them increases, so in this paper we describe quite a complex 'case study', which we think is a fairly typical of part of an industrial control system analysis. It is based on the use of a multivariable Smith predictor for a multivariable plant which contains time delays.