| Scheduling theory first appears in the mid 1950s. Since then the problems addressed become closer to industrial applications, thus increasing in complexity. The layout of the shops taken into account are closer and closer to those met in practice: we encounter shops where the machines are found in different multiple copies, shops where an operation may require several resources simultaneously, or with multipurpose machines, etc. At the same time the embedded constraints are more and more concrete: many authors take into account release dates, the preemption of the jobs, the resource availabilities, etc.
Paradoxically, the literature shows that in the majority of the problems addressed, schedules are only evaluated by a single criterion. During the diflFerent phases of planning different criteria can be considered. At a strategic level, at the long term planning phase with several years in view, the objectives concern minimising the costs related to the investment plans for materials, finance, or personel, related to the choice of new directions, or the launching of publicity campaigns. For tactical planning at the medium term phase with several months in view, the objectives always focus on minimising the costs: stock costs (supply or interruption of stocks), costs of getting supplies, costs of modifying production capacity, launching costs, costs of modifying production systems and certain commercial costs ([Merce, 1987], [Giard, 1988]).
At the short term planning phase (with the order of a week in view), or scheduling phase, several objectives require the attention of the production executive: above all he must consider the delays that satisfy the customer, next, he must minimise the work-in-process costs in the shop, and finally he must minimise the manufacturing costs related to the time spent to set up the machines or idle periods of the machines. Therefore, a scheduling problem involves multiple criteria. |