| Recent efforts in research and software development in the Manufacturing Production Planning field, have focused on enterprise-wide Supply Chain Management solutions (SCM). SCM aims at eliminating the most important shortfall of Manufacturing Resource Planning Systems (MRP). MRP relies on a sequential planning algorithm which fails to address simultaneously resource and flow constraints of end products and components. The MRP's backwards explosion of production requirements results in sub-optimal and inflexible solutions. In contrast, SCM Systems rely on concurrent optimization to create a globally superior schedule. As a result, the Enterprise problem size increases significantly. The common shortcut is the arbitrary and unstructured elimination or aggregation of important production system dynamics.; We propose a "bottom-up" approach that introduces Cellular Manufacturing as a means of natural decomposition of the system. Production Cells are defined as subsystems with common responsibilities (types of parts produced) or functionalities (types of resources owned). Decision support tools at each Production Cells use standard methods to define very precisely the best a-priori allocation of labor to resources and resources to part operations. This is based on novel Standard Work Design algorithms we propose and develop. After Production Cell Standard Work Design is performed, an enterprise-wide Production Planning algorithm is employed to synchronize cell-specific production requirements. Unlike most SCM approaches, which aggregate arbitrarily and ignore important constraints, the important characteristics of each Cell and its production and process idiosyncrasies are modeled explicitly at the Production Cell level and passed on to the Enterprise-wide algorithm. Thus, the resulting Cell specific production requirements are feasible and can be trusted by the Cell personnel. The Enterprise Level problem is solved using Linear Programming (LP) and provides sensitivity information that can be used to reallocate capacity across Cells. Alternative LP formulations and information exchange mechanisms are described and compared. Finally, each Cell tracks its production targets by employing detailed Scheduling algorithms.; We also compare against current industry practices (MRP) on industry size problems. Significant benefits in cycle time reduction and delivery performance are reported. In addition, a flexible enterprise data model is proposed for storing and exchanging information and software design considerations are addressed. |
