An approach to generate comprehensive piecewise linear interpolation of pareto outcomes to aid decision making
- 248 Downloads
Multiple criteria decision making is a well established field encompassing aspects of search for solutions and selection of solutions in presence of more than one conflicting objectives. In this paper, we discuss an approach aimed towards the latter. The decision maker is presented with a limited number of Pareto optimal outcomes and is required to identify regions of interest for further investigation. The inherent sparsity of the given Pareto optimal outcomes in high dimensional space makes it an arduous task for the decision maker. To address this problem, an existing line of thought in literature is to generate a set of approximated Pareto optimal outcomes using piecewise linear interpolation. We present an approach within this paradigm, but one that delivers a comprehensive linearly interpolated set as opposed to its subset delivered by existing methods. We illustrate the advantage in doing so in comparison to stricter non-dominance conditions imposed in existing PAreto INTerpolation method. The interpolated set of outcomes delivered by the proposed approach are non-dominated with respect to the given Pareto optimal outcomes, and additionally the interpolated outcomes along uniformly distributed reference directions are presented to the decision maker. The errors in the given interpolations are also estimated in order to further aid decision making by establishing confidence in achieving true Pareto outcomes in their vicinity. The proposed approach for interpolation is computationally less demanding (for higher number of objectives) and also further amenable to parallelization. We illustrate the performance of the approach using six well established tri-objective test problems and two real-life examples. The problems span different types of fronts, such as convex, concave, mixed, degenerate, highlighting the wide applicability of the approach.
KeywordsPareto front approximation Linear interpolation Decision making
The third author would like to acknowledge the support from Australian Research Council (ARC) Future Fellowship. Thanks also to Markus Hartikainen for discussions regarding PAINT.
- 4.Deb, K., Thiele, L., Laumanns, M., Zitzler, E.: Scalable multi-objective optimization test problems. In: Proceedings of the IEEE Congress on Evolutionary Computation. pp 825–830. IEEE, Honolulu, USA (2002)Google Scholar
- 6.Eyvindson, K., Hartikainen, M., Kurttila, M.: Towards constructing a Pareto front approximation for use in interactive forest management planning. In: Proceedings of the International Conference on Implementation of DSS Tools into the Forestry Practice. Technical University in Zvolen (2013)Google Scholar
- 9.Hartikainen, M., Miettinen, K.: A computationally inexpensive approach in multiobjective heat exchanger network synthesis. Proceedings of the International Conference on Applied Operational Research. 10, 99–109 (2010)Google Scholar
- 12.Hartikainen, M., Ojalehto, V.: Demonstrating the applicability of PAINT to computationally expensive real-life multiobjective optimization. arXiv preprint arXiv:1109.3411 (2011)
- 16.Kuhn, T., Ruzika, S.: A coverage-based Box-Algorithm to compute a representation for optimization problems with three objective functions. J. Global Optim. 1–20 (2016). doi: 10.1007/s10898-016-0425-5
- 17.Lotov, A.V., Miettinen, K.: Visualizing the Pareto frontier. In: Multiobjective optimization, Lecture Notes in Computer Science, vol 5252, pp 213–243. Springer (2008). doi: 10.1007/978-3-540-88908-3_9
- 20.Lovison, A., Hartikainen, M.: On generalizing Lipschitz global methods for multiobjective optimization. In: Proceedings of the International Conference on Evolutionary Multi-Criterion Optimization. Lecture Notes in Computer Science, vol 9019, pp 264–278 (2015)Google Scholar
- 23.Monz, M.: Pareto Navigation—Interactive Multiobjective Optimisation and Its Application in Radiotherapy Planning. Ph.D. thesis, Department of Mathematics, Technical University of Kaiserslautern (Germany) (2006)Google Scholar
- 28.Wiecek, M.M., Ehrgott, M., Engau, A.: Continuous multiobjective programming. In: Multiple Criteria Decision Analysis. International Series in Operations Research and Management Science, vol 233, pp 739–815. (2016). doi: 10.1007/978-1-4939-3094-4_18