A Model Based Interactive Graphical Scheduler

Advanced Process Combinatorics，Inc.提出了一种交互式调度系统，它将基于模型的优化与电子甘特图相结合。该系统(SMART/IGS)将是基于面向对象模型的优化调度系统SMART(Scheduling And Material Resources Timing)的新组成部分。因此，IGS将与现有的描述调度问题的语言RCSPec和现有的面向对象的混合整数程序优化器MILP集成在一起。IGS的主要定义特征将是电子甘特图与基于基本模型的优化算法的耦合。用户将能够使用图形用户界面向日程中添加各种约束。通过在甘特图中的任务上单击鼠标，用户将能够确定任务的执行时间或在序列中的位置。优化器将完成调度，围绕用户已经指定的部分构建优化的调度。除了用户友好性的明显优势外，这还为用户提供了对优化算法的详细控制。添加约束(部分时间表规范)将使优化问题变得更小，从而更容易解决。这种用户引导的问题简化可以被用作使困难问题变得可解的方法。这一点以及优化算法的进步和更快的计算机提供了希望，即利用严格的基于模型的优化的交互式调度将在未来几年内成为可交付的商业产品。

Advanced Process Combinatorics, Inc., proposes an interactive scheduling system which integrates model based optimization with an electronic Gantt chart. This system (SMART++/IGS) will be a new component for the object-oriented model based optimization scheduling system, SMART++ (Scheduling and Material Resource Timing). As such, IGS will be integrated with RCSPec, an existing language for describing scheduling problems, and MILP++, an existing object-oriented mixed integer program optimizer. The main defining feature of IGS will be the coupling of the electronic Gantt chart to the underlying model based optimization algorithms. Users will be able to add a wide variety of constraints to the schedule using the graphical user interface. By clicking the mouse on a task in the Gantt chart the user will be able to fix the execution time, or the place of a task in a sequence. The optimizer will complete the schedule, building an optimized schedule around the pieces already specified by the user. In addition to the obvious advantages of user friendliness, this provides the user with detailed control over the optimization algorithm. The addition of constraints (partial schedule specification) will make the optimization problem smaller and thus, easier to solve. Such user guided problem simplification may be exploited as a method of rendering difficult problems solvable. This and advances in optimization algorithms and faster computers offer hope that interactive scheduling which utilizes rigorous model based optimization will be a deliverable commercial product in the next few years.