All the models in this system are run by using the software ILOG OPL Studio 2.1.3. OPL Studio [1] is an integrated development environment for combinatorial optimization applications. It can be effectively used for constructing and solving linear programming, integer programming, and constraint programming models. The model codes are written in optimization programming language (OPL) developed by Pascal Van Hentenryck [9].
OPL Studio has several tools including CPLEX 6.5.3 (mathematical programming solver), SOLVER 4.4 (solver for constraint programming) and SCHEDULER 4.4 (a tool developed for constraint based scheduling). Once a model is constructed, after compilation, OPL Studio automatically detects the the problem type and determines the most convenient solver to solve it.
ILOG OPL Studio has a user friendly graphical environment (Figure 4). Unlike most other programming environments, the compilation errors are displayed clearly. By using menus, it is possible to change parameter settings (iteration limit, tolerance level, etc), search procedures (depth first search, limited discrepancy search, etc). The display of solution in graphical format is possible such as resource utilizations, and charts for resource constrained scheduling problems.
Another feature of OPL Studio is its ability to decide on the most efficient solver for a model. It is able to solve both mathematical programming and constraint programming models. Since in the proposed DSS, the solution is obtained by utilizing both constraint programming and mathematical programming, OPL Studio is very convenient single system.
In this course scheduling problem, the size of the course, student, instructor data is very large and it is not easy to manipulate the data without a database. At the same time, for model runs, data is retrieved and stored several times that necessitates a database. Since OPL Studio has database connection capability, the data can easily retrieved and stored by OPL Studio before and after running the models.
OPL Studio has a special tool called scheduler that contains efficient algorithms for solving the constraints of resource constrained scheduling problems. This option makes it possible to define and solve the constraints related to the resources and the activities very easily. Without using complicated mathematical expressions, it is possible to define constraints for resources, resource capacities, activities, resource requirements of activities as simple as writing a sentence in an ordinary language. Since the course scheduling problem is a type of resource constrained scheduling problem, the classroom types can be represented as resources with available capacities and the class meetings are the activities to be scheduled whose durations equal to length of class meetings.
All these features make ILOG OPL Studio the ideal tool to use in a DSS for course scheduling at Bilkent University.
