In today's challenging industrial contexts, decision makers need versatile tools to quickly acquire, synthesize, simulate and optimize several solutions that can be compared from multiple points of view and according to various performance criteria. In the context of limited-time multicriteria decision-making problems, we have developed a planning and scheduling framework based on several hybrid tools composed of mathematical models, dedicated heuristics, stochastic local search meta-heuristics and simulation models. This paper focuses on improving the efficiency and accuracy of one of the components of the framework by quickly and accurately computing eigenelements. The computational approach overcomes both eigenvalue and eigenvector ill-conditioning through an inexpensive robust iterative refinement scheme based on a Newton-Kantorovich method and a QR algorithm with an improved stopping test. A dynamic version of the graphical synthetic views for decision makers is also presented, where it is possible to follow the evolutions of several iteratively improved solutions by any meta-heuristic. The validation of the component is done on an actual, highly constrained scheduling problem. The schedules provided by the plant information system or by the framework are ranked and visualized on the basis of five criteria. The available time for the decision-making process, list of orders and configurations of the machines determine the decision-making process, which consists either in rapidly computing several schedules or in comparing iteratively optimized schedules to evaluate the gains/losses of the criteria when switching from one schedule to another. Our tool can be used whenever ranking several solutions with multiple criteria is required