Scheduler search goals are designed to be easy to use and suited to the more usual cases. However, they have limited scope and may not be the best solution for certain types of problems. While Scheduler search goals can help in prototyping a project, they do not prevent you from developing the best heuristic algorithms for your problem within the Solver environment.

A goal is used to define the search for a solution to a model. The model for which an instance of a goal will search for a solution is specified via the IloSolver API. That is, once an IloSolver has extracted a model, a goal is associated with the model with the function IloSolver::solve(const IloGoal) or (IloSolver::startNewSearch(const IloGoal)). For more details see the IBM ILOG Solver Reference Manual.

Scheduler provides four predefined functions that return a goal to assign start times to activities in a schedule. For best results with these goals, activities should have bounded durations and resource demands. If precedence constraints between activities have negative delays, solutions may be missed.

Scheduler provides several goals for selecting resources from alternative resource sets.