Optimizing the Double Inverted Pendulum′s Performance via the Uniform Neuro Multiobjective Genetic Algorithm

An inverted pendulum is a sensitive system of highly coupled parameters, in laboratories, it is popular for modelling nonlinear systems such as mechanisms and control systems, and also for optimizing programmes before those programmes are applied in real situations. This study aims to find the optimum input setting for a double inverted pendulum（DIP）, which requires an appropriate input to be able to stand and to achieve robust stability even when the system model is unknown. Such a DIP input could be widely applied in engineering fields for optimizing unknown systems with a limited budget. Previous studies have used various mathematical approaches to optimize settings for DIP, then have designed control algorithms or physical mathematical models.This study did not adopt a mathematical approach for the DIP controller because our DIP has five input parameters within its nondeterministic system model. This paper proposes a novel algorithm, named Uni Neuro, that integrates neural networks（NNs） and a uniform design（UD） in a model formed by input and response to the experimental data（metamodel）. We employed a hybrid UD multiobjective genetic algorithm（HUDMOGA） for obtaining the optimized setting input parameters. The UD was also embedded in the HUDMOGA for enriching the solution set, whereas each chromosome used for crossover, mutation, and generation of the UD was determined through a selection procedure and derived individually. Subsequently, we combined the Euclidean distance and Pareto front to improve the performance of the algorithm. Finally, DIP equipment was used to confirm the settings. The proposed algorithm can produce 9 alternative configured input parameter values to swing-up then standing in robust stability of the DIP from only 25 training data items and 20 optimized simulation results. In comparison to the full factorial design, this design can save considerable experiment time because the metamodel can be formed by only 25 experiments using the UD. Furthermore, the proposed algorithm can be applied to nonlinear systems with multiple constraints.

Design-space adaptation method for multiobjective and multidisciplinary optimization

This paper developed a new method that adaptively adjusts a design space by considering the actual solution distribution of a problem to overcome the conventional design-space adaptation method that assumes the solutions distribution to be a normal distribution because the distributions of solutions are rarely normal distributions for real-world problems.The developed method was applied to nineteen multiobjective test functions that are widely used to evaluate the characteristics and performance of optimization approaches.The results showed that this method adapted the design space to an appropriate design space where the solution existence probability was high.The optimization performance achieved using the developed method was higher than that of the conventional methods.Furthermore,the developed method was applied to the conceptual design of an unmanned spacecraft to confirm its validity in real-world design and multidisciplinaryoptimization problems.The results showed that the Pareto solutions of the developed method were superior to those of conventional methods.Additionally,the optimization efficiency with the developed method was improved by more than 1.4 times over that of the conventional methods.In this regard,the developed method has the potential to be applied to complicated real-world optimization problems to achieve better performance and efficiency.