Handling Uncertainties Through Reliability-Based Optimization Using Evolutionary Algorithms

Uncertainties in design variables and problem parameters are inevitable and must be considered in an optimization task, if reliable optimal solutions are to be found. Besides the sampling techniques, there exist a number of reliability-based probabilistic optimization techniques for systematically handling such uncertainties. In this paper, first we present a brief review of these classical probabilistic procedures. Thereafter, we discuss different optimization tasks in which these classical reliability-based optimization procedures will, in general, have difficulties in finding true optimal solutions. These probabilistic techniques are borrowed from classical literature and are extended to constitute efficient reliability-based single and multi-objective evolutionary algorithms for solving such difficult problems. Due to the global perspective of evolutionary algorithms, first, we demonstrate the proposed methodology is better able to solve reliability based optimization problems having multiple local-global solutions. Second, we suggest introducing an additional objective of maximizing the reliability index along with optimizing the usual objective function and find a number of Pareto-optimal solutions trading-off between the objective value and corresponding reliability index, thereby allowing the designers to find solutions corresponding to different reliability requirements for a better application. Finally, the concept of single-objective reliability-based optimization is extended to multi-objective optimization of finding a reliable frontier, instead of a single reliable solution. These optimization tasks are illustrated by solving a number of test problems and a well-studied automobile design problem. Results are also compared with a couple of standard classical reliability-based methodologies. This paper demonstrates how classical reliability-based concepts can be used in single and multi-objective evolutionary algorithms to enhance their scope in handling uncertainties, a matter which is common in most real-world problem solving tasks.

[1]  M. Rosenblatt.  Remarks on a Multivariate Transformation , 1952 .

[2]  John H. Holland,et al.  Adaptation in Natural and Artificial Systems: An Introductory Analysis with Applications to Biology, Control, and Artificial Intelligence , 1992 .

[3]  David E. Goldberg,et al.  Genetic Algorithms in Search Optimization and Machine Learning , 1988 .

[4]  A. Kiureghian,et al.  Optimization algorithms for structural reliability , 1991 .

[5]  Singiresu S Rao,et al.  Genetic algorithmic approach for multiobjective optimization of structures , 1993 .

[6]  Kalyanmoy Deb,et al.  Simulated Binary Crossover for Continuous Search Space , 1995, Complex Syst..

[7]  Henrik O. Madsen,et al.  Structural Reliability Methods , 1996 .

[8]  Thomas A. Cruse,et al.  Reliability-Based Mechanical Design , 1997 .

[9]  Kalyanmoy Deb,et al.  Time Scheduling of Transit Systems With Transfer Considerations Using Genetic Algorithms , 1998, Evolutionary Computation.

[10]  K. Deb.  An Efficient Constraint Handling Method for Genetic Algorithms , 2000 .

[11]  Xiaoping Du,et al.  A MOST PROBABLE POINT BASED METHOD FOR UNCERTAINTY ANALYSIS , 2000 .

[12]  Yun Li,et al.  Optimization and robustness for crashworthiness of side impact , 2001 .

[13]  Kalyanmoy Deb,et al.  A fast and elitist multiobjective genetic algorithm: NSGA-II , 2002, IEEE Trans. Evol. Comput..

[14]  Dhanesh Padmanabhan,et al.  RELIABILITY BASED OPTIMIZATION USING APPROXIMATIONS WITH APPLICATIONS TO MULTI-DISCIPLINARY SYSTEM DESIGN , 2002 .

[15]  Xiaoping Du,et al.  Sequential Optimization and Reliability Assessment Method for Efficient Probabilistic Design , 2004, DAC 2002.

[16]  Dhanesh Padmanabhan,et al.  Reliability-Based Optimization for Multidisciplinary System Design , 2003 .

[17]  Sankaran Mahadevan,et al.  Reliability-Based Design Optimization Methods , 2004, DAC 2004.

[18]  Wei Chen,et al.  An integrated framework for optimization under uncertainty using inverse Reliability strategy , 2004 .

[19]  J. Tu,et al.  A Single-Loop Method for Reliability-Based Design Optimization , 2004, DAC 2004.

[20]  Harish Agarwal,et al.  Reliability Based Design Optimization: Formulations and Methodologies , 2004 .

[21]  Kalyanmoy Deb,et al.  Evolutionary Multi-objective Environmental/Economic Dispatch: Stochastic Versus Deterministic Approaches , 2005, EMO.

[22]  Aravind Srinivasan,et al.  Innovization: innovating design principles through optimization , 2006, GECCO.

[23]  R. K. Ursem.  Multi-objective Optimization using Evolutionary Algorithms , 2009 .