I was taught at Toyota that there is one bottleneck. Once you improve that bottleneck so much as to no longer be the bottleneck, another bottleneck is already present to be improved upon if necessary. Even if all processes are able to meet the current demand, there is still a bottleneck but there is no need to improve the bottleneck since customer demand is being met. At that point our focus may turn to reduce cost per unit or to reduce inventory.
David Keown Nicely put. Thanks for posting. It sounds as though Toyota has a slightly different take than Goldratt. If I had to teach one of these, I'd rather teach Toyota's method than Goldratt's, but only because I think Toyota's would be easier for most people to understand.
You said in the end that E will remain bottleneck but you also said that A constraint can be called bottleneck but a bottleneck is not always a constraint. So E must be known as constraints. To my understanding there might be several bottlenecks in system but constraint has to be only one at any given time.
Hi there, your question may have different answers depending the state of the system, i.e. all machines and buffers empty or system running with inventories/buffers in between resources. It also depends if production is pushed or regulated via pull system, if products are all same or customized... In order to know the lead time, you must know the inventory/buffer level, if FIFO is applied and machine cycle time. I hope this helps.
