Optimization-based decision support systems for multiobjective decision making

决策作为人类最重要的活动之一，在当今复杂和快速变化的环境中变得越来越困难。在许多工业部门，管理人员不得不面对越来越复杂的问题，这些问题往往涉及多个不可分割的，往往相互竞争的目标的同时优化。因此，决策支持系统（DSS）在所有学科的商业智能组织决策中发挥着越来越重要的作用。此外，结合人工智能和运筹学技术来解决大型问题也有利于DSS的发展。建议的研究计划的主要目标是开发高效，灵活的优化工具作为DSS的主要组成部分，以解决工业问题。用于衡量解决方案策略的相关性和性能的主要应用是生产调度。 本研究有三个方向：（1）对复杂问题进行建模，特别是通过同时优化多个相互矛盾的目标，以丰富理论，同时开发有用的工具;（2）开发和实施高效，实用和专门设计的算法来解决困难的优化问题：自然启发的算法在这里受到青睐，因为它们在几十年的灵活性和性能方面的记录得到了证明;（3）推进计算方法，特别是在并行和分布式计算机体系结构上的基于优化的决策支持系统，以减少执行时间，同时提高解决方案的质量。*

Decision making, one of the most important human activities, is ever more difficult in today's complex and rapidly changing environment. In many industrial sectors, managers have to face increasingly complex problems which often involve the simultaneous optimization of multiple incommensurable and often competing objectives. Decision Support Systems (DSS) are therefore playing an ever more important role in organizational decision making for business intelligence in all disciplines. Moreover, combining artificial intelligence and operations research techniques to tackle large problems can also benefit the development of DSS.**The principal objective of the proposed research program is to develop efficient and flexible optimization tools as the main component of DSS in order to solve industrial problems. The principal application used to measure the pertinence and performance of solution strategies is production scheduling. This research has three orientations: (1) modeling complex problems, especially by the simultaneous optimization of several contradictory objectives, in order to enrich the theory but also to develop useful tools; (2) developing and implementing efficient, practical and specifically designed algorithms to solve difficult optimization problems: nature-inspired algorithms are favored here because of their proven record in flexibility and performance over several decades; (3) advancing computational methods, specifically for optimization-based decision support systems on parallel and distributed computer architectures, in order to obtain reductions in execution time but also to improve solution quality.************