Finding near-optimum and diverse solutions for a large-scale engineering design problem

Optimization problems in real-world usually involve a large number of variables and constraints. Moreover, practitioners also need to find a reasonable solution quickly, despite the fact that solution evaluation can be expensive. In addressing such optimization tasks, practitioners often have a completely different set of objectives than the usual trend of finding a single and accurate optimal solution. First, with the limited computational time availability and large dimension, it is not accepted (and in most occasions, not possible) to find the exact optimum solution of the problem, rather the focus must be spent to identify a near-optimal solution in a quick computational time. Second, instead of concentrating on a single solution, practitioners are often interested in evaluating a diverse set of solutions from the search space having similar objective values. Third, when a solution is finally chosen, practitioners are interested in exploring its vicinity to understand the trade-off between objective value and constraint violation, so that a more informed description of the problem can be ascertained. We highlight all these practicalities through a large-scale engineering design problem and present a suitable evolutionary-cum-classical, single-cum-multi-objective optimization method. The methods are generic and can be applied to other complex optimization tasks.

[1]  Yi Zhang,et al.  An "ageing" operator and its use in the highly constrained topological optimization of HVAC system design , 2005, GECCO '05.

[2]  Xin Yao,et al.  Stochastic ranking for constrained evolutionary optimization , 2000, IEEE Trans. Evol. Comput..

[3]  Marco Dorigo,et al.  Genetic Algorithms and Highly Constrained Problems: The Time-Table Case , 1990, PPSN.

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

[5]  Kalyanmoy Deb,et al.  An Evolutionary Many-Objective Optimization Algorithm Using Reference-Point-Based Nondominated Sorting Approach, Part I: Solving Problems With Box Constraints , 2014, IEEE Transactions on Evolutionary Computation.

[6]  J. Dennis,et al.  Mixed Variable Optimization of the Number and Composition of Heat Intercepts in a Thermal Insulation System , 2001 .

[7]  Thomas Bäck,et al.  Optimizing highly constrained truck loadings using a self-adaptive genetic algorithm , 2015, 2015 IEEE Congress on Evolutionary Computation (CEC).

[8]  Marzouk Benali,et al.  Multi-objective self-adaptive algorithm for highly constrained problems: Novel method and applications , 2010 .

[9]  Kalyanmoy Deb,et al.  An Evolutionary Many-Objective Optimization Algorithm Using Reference-Point Based Nondominated Sorting Approach, Part II: Handling Constraints and Extending to an Adaptive Approach , 2014, IEEE Transactions on Evolutionary Computation.

[10]  Roberto Piola Evolutionary solutions to a highly constrained combinatorial problem , 1994, Proceedings of the First IEEE Conference on Evolutionary Computation. IEEE World Congress on Computational Intelligence.

[11]  Kalyanmoy Deb,et al.  A Computationally Efficient Evolutionary Algorithm for Real-Parameter Optimization , 2002, Evolutionary Computation.