Genetic algorithm optimization of multi-peak problems: studies in convergence and robustness

doi:10.1016/0954-1810(95)95751-Q

Artificial Intelligence in Engineering

Volume 9, Issue 2, 1995, Pages 75-83

https://doi.org/10.1016/0954-1810(95)95751-Q Get rights and content

Abstract

Engineering design studies can often be cast in terms of optimization problems. However, for such an approach to be worthwhile, designers must be content that the optimization techniques employed are fast, accurate and robust. This paper describes recent studies of convergence and robustness problems found when applying genetic algorithms (GAs) to the constrained, multi-peak optimization problems often found in design. It poses a two-dimensional test problem which exhibits a number of features designed to cause difficulties with standard GAs and other optimizers. The application of the GA to this problem is then posed as a further, essentially recursive problem, where the control parameters of the GA must be chosen to give good performance on the test problem over a number of optimization attempts. This overarching problem is dealt with both by the GA and also by the technique of simulated annealing. It is shown that, with the appropriate choice of control parameters, sophisticated niche forming techniques can significantly improve the speed and performance of the GA for the original problem when combined with the simple rejection strategy commonly employed for handling constraints. More importantly, however, it also shows that more sophisticated multi-pass, constraint penalty functions, culled from the literature of classical optimization theory, can render such methods redundant, yielding good performance with traditional GA methods.

References (5)

J.N. Siddall
D.E. Goldberg

There are more references available in the full text version of this article.

Cited by (73)

Design of Fresnel acoustic reflector for sub-wavelength broadband sound diffusion
2024, Materials and Design
Sound diffusion performance is considered as an important issue in building acoustics. A well-diffused sound field can effectively enhance the human listening experience. In this context, inspired by Fresnel lenses in optics, we design a sub-wavelength acoustic metasurface (termed “Fresnel acoustic reflector”) to achieve broadband uniform diffuse reflections. To this end, an optimization strategy that combines Monte-Carlo search and other derivative-free algorithms is proposed to ensure, as far as possible, a global optimum for the objective function when dealing with the multi-peak problem encountered in complex sound fields. Numerical simulations show that the average normalized diffusion coefficient of the designed Fresnel reflector can reach up to 0.6 in the sub-wavelength frequency range 500–3400 Hz, which exhibits a better broadband diffusion performance compared to the traditional Schroeder diffusers. The simulated sound pressure field is also verified by experimental measurements. From a general perspective, the proposed optimization strategy and simulation framework can benefit the design of interior components in building acoustics.
Evolutionary ORB-based model with protective closing strategies
2021, Knowledge-Based Systems
Citation Excerpt :
However, an uneven solution space with multiple peaks cannot guarantee a gradient indicating the correct direction. GA is a heuristic approach in evolutionary computation [22], which has proven highly effective in nonconvex multi-peak optimization problems [23]. This approach is based on the concept of survival by natural selection [24], in which a strong individual (a strong parameter set) has a higher likelihood of survival.
Opening range breakout (ORB) is a well-known intraday trading strategy via technical analysis. ORB lacks robustness against market uncertainties (e.g., information from contradictory sources), and does not consider all relevant market characteristics. Furthermore, the closing strategies in generic ORB are not well defined. In this study, we developed an evolutionary ORB-based model, which utilized historical data to optimize thresholds in order to enhance profitability, and developed protective closing strategies aimed at to prevent unacceptable losses. Selecting appropriate thresholds and parameters for ORB is a non-trivial task, due to the fact that the search space exceeds sixty-five thousand options. We used evolutionary computation to derive rational strategies and parameters for ORB. The proposed framework based on a genetic algorithm optimizes the parameters related to threshold selection and protective closing strategies. In experiments, this resulted in annual returns of 9.3% (representing a 2.8% improvement over the original strategy) and Sharpe ratio of 2.5 (an improvement of 1.0), while reducing the maximum drawdown by half. The proposed scheme also reduced computational overhead by 89% compared to a grid search.
A meta optimisation analysis of particle swarm optimisation velocity update equations for watershed management learning
2018, Applied Soft Computing
Citation Excerpt :
The field of meta optimisation dates back to 1978 when it was first applied to tune a genetic algorithm [32]. In the years since, meta optimisation has been applied to ant colony optimisation [6], differential evolution [34], COMPLEX-RF [22], particle swarm optimisation [31,35] and genetic algorithms [14,4,20]. There are limitations to the previous studies conducted in applying meta optimisation to PSO.
Particle swarm optimisation (PSO) is a general purpose optimisation algorithm used to address hard optimisation problems. The algorithm operates as a result of a number of particles converging on what is hoped to be the best solution. How the particles move through the problem space is therefore critical to the success of the algorithm. This study utilises meta optimisation to compare a number of velocity update equations to determine which features of each are of benefit to the algorithm. A number of hybrid velocity update equations are proposed based on other high performing velocity update equations. This research also presents a novel application of PSO to train a neural network function approximator to address the watershed management problem. It is found that the standard PSO with a linearly changing inertia, the proposed hybrid Attractive Repulsive PSO with avoidance of worst locations (AR PSOAWL) and Adaptive Velocity PSO (AV PSO) provide the best performance overall. The results presented in this paper also reveal that commonly used PSO parameters do not provide the best performance. Increasing and negative inertia values were found to perform better.
Meta-harmony search algorithm for the vehicle routing problem with time windows
2015, Information Sciences
Citation Excerpt :
The main role of the meta-HSA optimizer is to adjust HSA parameter values, the local search algorithm type and the local search configurations (parameter values and the neighborhood structures) during the search without any external influence. The main difference between the proposed meta-HSA and the existing ones [24–28] is that the proposed meta-HSA is used to adjust the parameter values, local search types and local search configurations (parameter values and neighborhood operators) while the existing meta-optimizers only adjust the parameter values. Furthermore, the proposed meta-HSA adjusts these components and configurations in an online manner, while existing ones use training and testing instances, which might make them well suited to the training instances only.
It has been proven that the hybridization of the harmony search algorithm with a local search algorithm (LS) is essential to compensate for the inadequacy of its exploitation. However, the success of this hybridization relies on the achievement of a proper balance between HSA exploration and LS exploitation. The question is can we obtain this balance by adaptively selecting (i) HSA parameter values, (ii) LS and its parameters and (iii) the LS neighborhood structures? To address these issues, this work proposes a meta-harmony search algorithm (meta-HSA) that uses two HSA algorithms, an HSA-optimizer and HSA-solver. The HSA-optimizer will adaptively adjust the components and the configurations of the HSA-solver based on the search status. The HSA-solver, which is a hybridization of HSA and LS, takes the configuration generated by the HSA-optimizer as input and then solves the given problem instance. That is, the HSA-optimizer operates on the components and the configurations of the HSA-solver, while the HSA-solver operates directly on the given problem instance (the solution search space). The proposed meta-HSA was applied to Solomon’s vehicle routing problem with time windows benchmark to verify its effectiveness compared with standard HSA and the state-of-the art methods. The results of the comparison confirmed that the meta-HSA produces competitive results with respect to the other methods. Therefore, we can conclude that the meta-optimization technique does assist the hybrid HSA in obtaining the appropriate selection of its components and configurations during the search process. This demonstrates that the meta-HSA can provide a proper balance between exploration and exploitation by adaptively selecting (i) HSA parameter values, (ii) LS and its parameters and (iii) the LS neighborhood structures. Moreover, the meta-HSA optimizer decreases the effort exerted by the user in tuning these components and configurations.
Field performance of a genetic algorithm in the settlement prediction of a thick soft clay deposit in the southern part of the Korean peninsula
2015, Engineering Geology
Citation Excerpt :
A back-analysis method based on a genetic algorithm (GA) can be used as a parallel and global search tool that emulates natural genetic operators. GAs generally show better performance when searching for a solution than conventional optimization algorithms because GAs, which make use of an entire set of solutions spread throughout the solution space, are less affected by local optima (Holland, 1975; Goldberg, 1989; Keane, 1995). Park et al. (2009) showed that the GA back-analysis method has the advantage of robustly searching for a global solution while avoiding local solutions compared with conventional optimization schemes in a multi-dimensional consolidation problem with three consolidation layers.
Long-term settlement data of a thick soft clay deposit improved by vertical drains frequently shows different field settlement behaviors from laboratory-driven predictions by conventional theories due to the variability and uncertainty of the soil properties, modeling simplifications, and types of prediction methods. This paper presents the field application of a back-analysis method based on a genetic algorithm (GA) to evaluate the performance of a new settlement prediction method compared with conventional graphical settlement prediction methods, such as the hyperbolic method and the Asaoka method. The GA back-analysis method shows better flexibility in modifying surcharging plans and adaptability to multi-layered thick soft soil deposits at the early stages of post-construction settlement. Thus, this new settlement prediction method enables geotechnical engineers to subsequently modify the heights of surcharge fills in accordance with field settlement data monitored in the interim for rapid and cost-effective construction. The comparative results show that the GA back-analysis method is capable of superior field performance in settlement predictions compared with two conventional graphical methods, within a margin of less than 200 mm in a thick soft clay deposit with multiple layers under complex loading conditions.
Designing and optimizing deviated wellbore trajectories using novel particle swarm algorithms
2014, Journal of Natural Gas Science and Engineering
Wellbore trajectory design is a determinant issue in drilling engineering. This paper introduces a new stochastic approach for drilling trajectory design applying continuous particle swarm algorithms to find the optimum drilling measured depth of directional and horizontal wells in 3-D space. Considering all the constraints and limitations, the final goal is to determine all geometrical well parameters, in order to achieve the optimum measured depth to the desired target location. Particle swarm optimization is a computational algorithm inspired from natural behavior of some animal societies, such as flocks of birds and shoals of fish. In this review, the trajectory design method for an objective function, originally proposed by Adams and Charrier (1985), is explored and developed. Also, the attributes of the particle swarm optimization (e.g., Onwunalu, 2010) and Meta-optimization (e.g., Pedersen, 2010a, Pedersen, 2010b) algorithms are considered and compared. These algorithms are then applied to demonstrate the determination of true measured depth of example horizontal wellbores as the objective function. Faster convergences, better final points which satisfy all constraints imposed on the drilling paths and population diversity maintenance to help the algorithms find better solutions, are positive characteristics of the solutions found using the algorithms proposed. These algorithms make promising new tools for designing economically-effective trajectories for deviated wells.
MATLAB codes for the PSO algorithms evaluated are provided as appendices to this article.

View all citing articles on Scopus

View full text