Supply chain management is the arena consisting of material & services procurements, production planning, distribution planning & relevant transportation management. In all of these areas, advance planning is done so that men, material & machines can be utilized efficiently. For example, the production shop should plan in advance from the data available about sales orders & forecasts. The procurement department should plan about procuring materials & services from the sales orders & forecasts. Likewise all departments need planning so that required resources are available at right time and thus delays can be minimized.
Making planning for these departments is very difficult though. the reason is that resources are finite and the same task can be fulfilled using many alternative ways. At the same time making sure that required resources are available at right time is a difficult task to plan. For example, suppose you have to make a plan to load a machine optimally given the constraints of set up time, production time, routing etc. You have many such machines and you make products on the same type of machines. The production plan should be such that it should minimize set up time and maximize machine usage to the maximum. At the same time the capacity constraint should also be taken care of. For example the machine should be loaded only for 8 hours per day and 6 days a week. The required quantities of raw material or the intermediate material should also be available at the time when the machine is loaded.
There could be so many constraints for any production or distribution planning that it is realistically impossible to make the perfect plan. Given the practical limitations, how anyone can make such plans for large production plants or distribution centers? Using excel sheets or simple planning tools one can make simple plans. But when complexity increases, it becomes difficult.
Some artificial intelligence tools can be used for supply chain management planning. Here some algorithms based on heuristics, artificial neural networks or any such suitable algorithm can be deployed.
Some software developers have used some such tools in making such plans. At universities research is going on to make better algorithms for such usage.