The research relates to the modelling and control of multi-stage production - inventory systems in high volume low-mix manufacturing industry. Examples of such class include typically the automotive and white goods industries. The research has included the investigation of available mathematical control techniques in the "optimal" control of manufacturing systems, a study of their applicability and the practical implications of their use in a manufacturing environment. Earlier work in the field of multivariable control theory has shown the potential of application in industrial management. In this thesis, previous work is extended whereby explicit consideration is given to some practical constraints existing in a typical manufacturing environment. It is considered that the research carried out has contributed to the development of multivariable control theory as applied to practical control problems with constraints. This has been achieved by the use of structured canonical forms and the exploitation of their particular ordered properties, resulting in the development of practical control models. The automotive industry has been used as a practical case study and modelled as a linear discrete-time control problem. The models have been developed in close liaison with a car manufacturing company in the U.K., and have been shown to produce practical control policies in the areas of both capacity requirements planning and inter-stage float levels. Particular attention is given to existing practical constraints of such systems. The approach is extended to deal effectively with a more general multi-product environment. It is noted that multi-product environment is of a more complex nature than single product since it involves the consideration of competition for the limited resources that have to be shared out "sub-optimally", in addition to providing smooth control of the responses. The development and application of multivariable control theory as described in this thesis is shown to provide an effective methodology for the solution of dynamic production control problems of multi-stage production-inventory systems in both single and multi-product environments.
|Date of Award