IBM ILOG Scheduler User's Manual > Integrated Applications > Incremental Scheduling: Using Scheduler with Multiple Problem Sub-models > Designing the Models > Modeling the Initial Problem

The initial problem is similar to the job shop problems we have examined in previous chapters. The differences are the number of activities and the use of discrete resources with capacity 3.

Therefore, the modeling of the original problem is quite straightforward. Users are referred to the function CreateOriginalModel and to Scheduling with Unary Resources: the Job-Shop Problem where the modeling of a job shop problem is described.

When a new job is created it is added to the initial model. Our assumption is that this model (the original model plus the new job) is too large to be solved as a single problem given the time and/or memory limitations of a real application. However, this does not prevent us from having a model of the entire problem because:

1) We do not solve this model as a whole.

2) A model in Concert Technology has a significantly lighter memory consumption than the same model extracted by a solver.

Therefore, once we have a solution to the original problem (see Solving the Original Model) and have added a new job to the original model, our task is to design secondary models that can be extracted and solved with a small memory and CPU time impact. The solutions to these submodels will then be efficiently combined into a solution for the overall problem. Thus, without actually having to extract and solve the overall problem, we will form a solution for it.

Each time we add a new job, the overall model and the solution to it (found via the combination of sub-model solutions) is changed. We will refer to the overall model, therefore, as the evolving model in that at any time it contains the model of the original problem plus whatever new jobs have been added: it has evolved over time. Similarly, we will refer to the solution to the evolving model as the evolving solution.