Group Technology (GT) is a manufacturing philosophy in the world that identify and classify part-part into groups of similar parts (part family) by exploiting the similarity in terms of product design and fabrication processes in the manufacturing cycle. GT aims to reduce setup time, material handling activities, throughput time, in-process inventory, space requirements, time idle machines, and control complexity, which in turn will increase production efficiency.
Cellular manufacturing (CM) is one application of GT. The underlying idea of CM is grouping machines into cells to produce a part family, part of a group that requires manufacturing processes are similar. CM decompose a system of production into several sub-systems, called the cell machinery (machine cell), where in each cell can be processed one or more parts without doing a full family of transfer between cells.
The first issue that must be addressed in the design of cellular manufacturing systems is the establishment phase of the cell machinery (machine cell formation). This stage is critical because the success of this system is highly dependent on the initial grouping of machines and parts. Absolute independence between the cell’s machinery is difficult to achieve in the real world, because there are some parts that must be processed in more than one cell. Part is called exceptional parts and machines that process is called bottleneck machine, both of which resulted in the transfer of inventory between cells that increase work-in-process, material handling costs, and lower system efficiency.
Switching between the cell can be eliminated by duplicating on a bottleneck machine or perform sub-contracts for the exceptional part. The process of utilizing the machine cell formation connectivity matrix (incidence matrix) machine-component, which is formed from the information contained on the part routing sheet.
The important thing to consider in forming machine cells are to maximize the level of independence between cells and maximizing machine utilization level in each cell without violating limits cell size and number of cells that have been set. Objective function (cost function) minimization of the number of switching parts between cells or minimization of duplication of bottleneck machines need to be properly designed and simple to meet the above matters without the need for large computing time. Machine cell formation problem is a form of data clustering problem.
Since the input data in the form of a matrix of connectedness machine-component binary valued (0 and 1), then it is classified as combinatorial optimization problems, ie search the optimum value of a function of discrete variables, which requires completion time increases along with the non-polynomial increasing the size of the problem. Many problems of theoretical and practical combinatorial optimization can only be solved by approximation methods (heuristic) by today’s computers.