Optimised Production Technology (OPT) is based on identifying bottlenecks in the production process - points at which items are processed at a slower rate - and maximising their output. OPT favours focusing on the shop floor, recognising that bottlenecks impact on the whole factory, then identifying the bottlenecks and aiming to optimise their use.
Consider a group of walkers. If they are to finish their walk at the same time, the maximum speed at which they can walk is the speed of the slowest. If the others accelerate, the slow walker is left behind and they have to wait for him to catch up.
Similarly, non-bottleneck operations that precede a bottleneck must produce items at the same rate as the bottleneck; otherwise, inventory will build up between the operations. OPT recommends producing only what is needed then balancing the flow of products through the factory.
It follows from this that utilisation of a resource - where one does as much work as can be done - is not the same as activation - where one does only what is needed. It also follows that saving time at a non-bottleneck does not reduce the overall lead time: it merely increases that workstation's idle time.
OPT recommends that the size of the transfer batch not be equal to that of the process batch: if part of a batch is passed to the next workstation before the rest, that workstation can start work on the batch earlier, resulting in reduced lead times. OPT also recommends that the size of the process batch be variable; it should be larger at bottlenecks in order to maximise output, and smaller elsewhere in order to reduce lead times and inventory.
Another OPT principle states that priority and capacity should be considered simultaneously. Under Materials Requirements Planning (MRP), planning lead times are decided, then priorities are assigned to jobs; the plan is then checked against capacity. OPT, in contrast, believes there is an important interaction between priority and capacity and that lead times are not fixed.
Finally, OPT states that the sum of the local optima is not equal to the optimum of the whole.
[Contents][Previous Page][Next Page]