In this paper we examine the problems of management and control of large scale multi-product multi-line batch manufacturing outside the mechanical engineering industries. The main examples of such non-job shop type manufacturing arises in the making of frequently purchased products e.g. small consumer goods, processed foods, cigarettes, candy, etc. The decision making problems which arise in such systems are extremely complex and the decision support tools for helping management at various levels could be of significant advantage in improving the productivity of such businesses. This paper describes a number of decision support tools which could be used at the different levels of the production hierarchy.The production hierarchy in virtually all manufacturing can be conveniently divided into 4 levels each of which is characterised by a significantly different time scale from that of the other levels. At the highest level, the time horizon of interest is 1 to 5 years and the decision making is of a strategic nature. At the next level down where the time horizon of interest is from 1-2 months to 1 year, the decision making is operational/tactical. The third level down concentrates on issues of production planning and scheduling and its horizon of interest varies from 1 hour to about 20 weeks. The lowest level deals with disturbances occurring over a very short horizon (a few seconds to about 1 hour) and at this level, most of the decision making is automatic (PID controllers, optimal control, self tuning regulators, etc.). Each higher level provides performance targets for the next level down.In this paper we examine the main issues which arise at the top three levels in the hierarchy and describe some decision support tools which have been developed by the author and others. At the top two levels the tools are designed to cope with the issues of competition whilst at the lower level (production planning and scheduling) the problems of allocating tasks to limited resources in the most cost effective way are addressed. Overall, we show how improved efficiency at each level can lead to significant improvements in productivity.