Multiobjective Meta-Heuristics in R&D Management

Large amounts of resources are at stake in research and development (R&D) management. Thus, particularly the selection of the "best" R&D projects provides a challenge for many organizations. The difficulties of this decision to a great extent arise from the multiple objectives that have to be taken into consideration. In addition, managers usually prefer interactive decision support which implies the need to determine the set of the (Pareto-) efficient project portfolios. While a brute-force complete enumeration procedure can be used for this purpose for quite small problems only, multiobjective meta-heuristic approaches provide an especially promising alternative. Accordingly, the project at hand combines the two fields of interactive decision support in R&D management and the newly arising area of multiobjective meta-heuristics. The project`s overall goal is to develop and/or extend and compare (new) multiobjective meta-heuristics with a focus on Tabu Search and Ant Colony Optimization. In order to achieve this goal multiobjective mathematical programming is used to model the underlying R&D project selection problem and computer science and computational analysis, respectively, are needed to implement, test, and analyze the performance of different meta- heuristics for the given problem class. However, the results of the project will also be of value in a variety of different decision problems (e.g., in transportation and logistics, healthcare management, or finance). The acquired know-how correspondingly will contribute to research in Austria and, furthermore, strengthen the academic capabilities to offer recent methods to Austrian companies and thus allow them to gain competitive advantages. Finally, the intended long-term cooperation with Professor Sun from the University of Texas should be a critical asset for future research projects on an international level.