|
|
Basic Terminology
What is a Schedule?
A schedule is an assignment of manufacturing tasks to given pieces of equipment at given times.
What is a Feasible Schedule?
A feasible schedule is one that satisfies all manufacturing constraints, including material balances, hard demands, inventory and resource limitations and user defined criteria such as campaign length and minimum run length. The word "feasible" is often misused in sales and marketing contexts. A schedule is not feasible if even a single constraint is violated. Often, one encounters phrases such as "Schedule X is more feasible than Schedule Y." One schedule may violate more constraints than another, but a given schedule can only be feasible or infeasible.
What is an Optimal Schedule?
The word "optimal" is a mathematical term with a very specific meaning and is often misused in sales and marketing contexts. In the context of scheduling, an optimal schedule is one that is feasible and has the best possible rating with respect to a performance criterion. Common performance criteria are maximizing profit or minimizing production costs. Many sales presentations and accompanying literature incorrectly use this term. Other than for very specific types of scheduling problems, no system can routinely generate optimal schedules. In fact, most systems do not even have the following three prerequisites for finding optimal schedules:
- A modeling structure which is capable of containing an optimal schedule.
- An algorithm capable of performing a set of operations which can yield an optimal schedule from the model.
- A means for proving that no other schedule will have a better rating.
Although several special purpose algorithms exist for highly specific classes of scheduling problems, generally speaking, the only way to guarantee that a schedule is optimal is to use mathematical programming techniques that provide a full model, engineered algorithms, and proof of schedule quality.
What is the difference between scheduling and planning?
Planning and scheduling both refer to the allocation of resources to meet product demands subject to process constraints. By convention, planning addresses a medium term time scale of months to a few years, while scheduling addresses a shorter time scale of days to weeks. In order to achieve computational savings, planning activities usually do not consider process details that are important for scheduling. However, this can often result in plans that are either too aggressive to implement in practice or so conservative that a facility will be underutilized. Ideally, one would like to use the most detailed process model possible for planning, which in the extreme would be a very large scheduling model. A very robust solution method would be required for this purpose.
|
COMPANY
