Combining Aggregation with Pareto Optimization: A Case Study in Evolutionary Molecular Design
This paper is motivated by problem scenarios in automated drug design. It discusses a modeling approach for design optimization problems with many criteria that can be partitioned into objectives and fuzzy constraints. The purpose of this remodeling is to transform the original criteria such that, when using them in an evolutionary search method, a good view on the trade-off between the different objectives and the satisfaction of constraints is obtained.
Instead of reducing a many objective problem to a single-objective problem, it is proposed to reduce it to a multi-objective optimization problem with a low number of objectives, for which the visualization of the Pareto front is still possible and the size of a high-resolution approximation set is affordable. For design problems where it is reasonable to combine certain objectives and/or constraints into logical groups by means of desirability indexes, this method will yield good trade-off results with reduced computational effort. The proposed methodology is evaluated in a case-study on automated drug design where we aim to find molecular structures that could serve as estrogen receptor antagonists.
Unable to display preview. Download preview PDF.
- 2.Branke, J., Deb, K.: Integrating user-preferences into evolutionary multi-objective optimization. In: Jin, Y. (ed.) Knowledge integration into evolutionary multiobjective optimization, Berlin, pp. 461–478 (2004)Google Scholar
- 3.Cvetkovic, D., Parmee, I.C.: Preferences and their application in evolutionary multiobjective optimization. IEEE TEC (6), 1, 42–57 (2002)Google Scholar
- 5.Derringer, G., Suich, R.: Simultaneous optimization of several response variables. Journal of Quality Technology 12, 214–219 (1980)Google Scholar
- 6.Harrington, E.C.: The desirability function. Industrial Quality Control 21, 494–498 (1965)Google Scholar
- 7.Inuiguchi, M., Ichihashi, H., Tanaka, H.: Fuzzy programming: A survey of Recent Developments. Stochastic versus Fuzzy Approaches to Multiobjective Mathematical Programming under Uncertainty, 45–68 (1990)Google Scholar
- 9.Laumanns, M., Thiele, L., Zitzler, E.: Archiving with guaranteed convergence and diversity in multi-objective optimization. In: Proceedings of the Genetic and Evolutionary Computation Conference, pp. 439–447 (2002)Google Scholar
- 13.Kruisselbrink, J.W., Bäck, T., van der Horst, E., IJzerman, A.P.: Evolutionary algorithms for automated drug design towards target molecule properties. In: Proceedings of the Conference on Genetic and Evolutionary Computation, pp. 1555–1562 (2008)Google Scholar
- 14.Trautmann, H., Mehnen, J.: A method for including a-priori preferences in multicriteria optimization. TR 49/2005, University of Dortmund, Germany, SFB 475 (2005)Google Scholar
- 15.Trautmann, H., Mehnen, J.: Integration of Expert’s Preferences in Pareto Optimization by Desirability Function Techniques. In: CIRP ICME 2006, Ischia, Italy, pp. 293–298 (2006)Google Scholar
- 16.Nicolaou, C.A., Pattichis, C.S.: Multi-objective de novo drug design using evolutionary graphs. Chemistry Central Journal 2008 2(suppl. 1), 7 (2007)Google Scholar
- 17.Trautmann, H., Weihs, C.: Pareto-Optimality and Desirability Indices. Department of Statistics, University of Dortmund (2004)Google Scholar
- 18.SciTegic, Inc. 10188 Telesis Court, Suite 100, San Diego, CA 92121, USA., http://www.scitegic.com/products_services/pipeline_pilot.htm
- 19.Sharma, B., Parmee, I.C., Whittaker, M., Sedwell, A.: Drug discovery: exploring the utility of cluster oriented genetic algorithms in virtual library design. In: Congress on Evolutionary Computation, pp. 668–675 (2005)Google Scholar