Inventory control involves the actual control of inventory; this can mean inventory of raw materials, works-in-progress or finished goods. Regardless of the type of inventory in question, inventory requires storage, and there is always a cost associated with that storage. Therefore, inventory control theory is concerned with all actions related to the storing of items and the consequences, both positive and negative, thereof.
One of the most common applications of inventory control theory is in the determination of the optimal quantity of inventory to be held. There are several mathematical models in use that can act as a useful tool in inventory control. These models strive to balance storage costs with order costs; the costs of shortages is also considered. While inventory control theory tends to be a bit shortsighted regarding the non-monetary costs of storage, and it makes assumptions regarding future demand and delivery that could not be known, inventory control theory is still a cost-saving tool, and is considered part of good business practice in manufacturing environments.
Deterministic models are a type of inventory control model that is used to determine the optimal inventory of a single item when demand is largely unknown. An example of this type of model is the classical lot size model, which assumes that demand is constant and continuous and that there is no delivery time and no shortage costs. When this model is used, inventory is built up at a continuous rate to meet a specified, or assumed, demand. An example of this would be a contract received in January to 100 model trains ready for November/holiday shopping. The deadline is 10 months away so the trains are produced at a rate of ten per month.
There also exist stochastic one-item models. These are used when demand is unknown. Stochastic models tend to be more realistic, and thus more applicable, because they consider the cost of shortages, the cost of ordering and the cost of storing and seek to develop an optimal inventory plan. While several models exist, many people will simply use trial-and-error or solve-for functions in spreadsheet software.
Multi-item models of inventory control look at several items of inventory at once. This is very useful because there are frequently cost-savings in an order of several items. Also, this type of model more accurately reflects the true inventory scenarios faced by manufacturing companies; it is rare to find a company that produces only one item with no variation. Multi-item models consider no only the storage, shortage and overage costs of inventory, but also the savings found in the common use of certain raw materials.
Forecasting is a necessary assumption in all inventory control. Without estimating customer demand, inventory cannot be controlled and shortage costs are practically inevitable. The biggest hurdle faced with forecasting is that forecasting is predominately subjective in nature; assumptions are drawn, but forecasts rarely involve quantitative fact. Forecasting models are distinguished by the time frame they forecast: immediate (up to one day; used in production with precise expiry, such as a bakery); short-term (up to one month; the most common, generally involves simple weighted averages for growth and seasonality); medium-term (up to one year; involves time series analyses and regression); and, long-term (up to 10 years; the province of think tanks and advanced regression analysis).
Many companies will make the mistake of using short-term forecasting methods to predict longer-term growth. Weighted averages are simply not very accurate in predicting demand that far out. While math may not be a strong suit, there are many resources, be it online, through mentoring (i.e. SBA, SCORE) or by the help of a friend, that will help an operations manager use the more appropriate time series and regression analyses typical of long-term forecasting, which will make the use of the inventory control models that much more accurate.