305,505 research outputs found
A Scalarizing One-Stage Algorithm for Efficient Multi-Objective Optimization
No full textA novel kriging-assisted algorithm is proposed for computationally expensive multi-objective optimization problems, such as those which arise in electromagnetic design. The algorithm combines the multiple objectives into a single objective, which it then optimizes using a one-stage method from singleobjective optimization. Its efficiency is demonstrated by comparison to a random search on a difficult test function
Scalarizing cost-effective multiobjective optimization algorithms made possible with kriging
No full textThe use of kriging in cost-effective single-objective optimization is well established, and a wide variety of different criteria now exist for selecting design vectors to evaluate in the search for the global minimum. Additionly, a large number of methods exist for transforming a multi-objective optimization problem to a single-objective problem. With these two facts in mind, this paper discusses the range of kriging assisted algorithms which are possible (and which remain to be explored) for cost-effective multi-objective optimization
Balancing Exploration and Exploitation using Kriging Surrogate Models in Electromagnetic Design Optimization
No full textThe balance between exploration and exploitation is an important issue when attempting to find the global minimum of an objective function. This paper describes how this balance may be carefully controlled when using Kriging surrogate models to approximate the objective function
Probability of improvement methods for constrained multi-objective optimization
No full textThis paper shows how the simultaneous consideration of multiple Kriging models can lead to useful metrics for the selection of design vectors in constrained multiobjective optimization. The savings in computational cost with such methods make them particularly useful for optimal electromagnetic design
An Enhanced Probability of Improvement Utility Function for Locating Pareto Optimal Solutions
No full textThis paper describes a novel utility function for choosing design vectors to evaluate in multi-objective optimization problems which are statistically most probable to be Pareto-optimal, given the points already evaluated. The method is tunable to the number of existing Pareto-optimal solutions that an unevaluated design vector is sought to dominate, is naturally parallelized, and removes any need for combining the multiple objectives into a single objective with a scalarizing function
Kriging methods for constrained multi-objective electromagnetic design optimization
No full textEThOS - Electronic Theses Online ServiceGBUnited Kingdo
Considerations of Accuracy and Uncertainty with Kriging Surrogate Models in Single-Objective Electromagnetic Design Optimization
No full textThe high computational cost of evaluating objective functions in electromagnetic optimal design problems necessitates the use of cost-effective techniques. This paper discusses the use of one popular technique, surrogate modelling, with emphasis placed on the importance of considering both the accuracy of, and uncertainty in, the surrogate model. After briefly reviewing how such considerations have been made in the single-objective optimization of electromagnetic devices, their use with kriging surrogate models is investigated. Traditionally, space-filling experimental designs are used to construct the initial kriging model, with the aim to maximize the accuracy of the initial surrogate model, from which the optimization search will start. Utility functions, which balance the predictions made by this model with its uncertainty, are often used to select the next point to be evaluated. In this paper, the performances of several different utility functions are examined using experimental designs which yield initial kriging models of varying degrees of accuracy. It is found that no advantage is necessarily achieved through searching for optima using utility functions on initial kriging models of higher accuracy, and that a reduction in the total number of objective function evaluations may be achieved by starting the iterative optimization search earlier with utility functions on kriging models of lower accuracy. The implications for electromagnetic optimal design are discussed
The consideration of surrogate model accuracy in single-objective electromagnetic design optimization
No full textThe computational cost of evaluating the objective function in electromagnetic optimal design problems necessitates the use of cost-effective techniques. This paper describes how one popular technique, surrogate modelling, has been used in the single-objective optimization of electromagnetic devices. Three different types of surrogate model are considered, namely polynomial approximation, artificial neural networks and kriging. The importance of considering surrogate model accuracy is emphasised, and techniques used to improve accuracy for each type of model are discussed. Developments in this area outside the field of electromagnetic design optimization are also mentioned. It is concluded that surrogate model accuracy is an important factor which should be considered during an optimization search, and that developments have been made elsewhere in this area which are yet to be implemented in electromagnetic design optimization
A hybrid one-then-two stage algorithm for computationally expensive electromagnetic design optimization
No full textA novel kriging-assisted algorithm is proposed for computationally expensive single-objective optimization. The principle behind the algorithm is to use information about objective function space at the earliest possible opportunity. After constructing a very small experimental design, a one-stage optimization algorithm is used to select further points to evaluate in design variable space. These points are then used in lieu of a traditional space-filling experimental design to construct the initial kriging model for a normal two-stage optimization algorithm
Going Beyond Counting First Authors in Author Co-citation Analysis
No full textThe present study examines one of the fundamental aspects of author co-citation analysis (ACA) - the way co-citation
counts are defined. Co-citation counting provides the data on which all subsequent statistical analyses and mappings
are based, and we compare ACA results based on two different types of co-citation counting - the traditional type that
only counts the first one among a cited work's authors on the one hand and a non-traditional type that takes into
account the first 5 authors of a cited work on the other hand. Results indicate that the picture produced through this non-traditional author co-citation counting contains more coherent author groups and is therefore considerably clearer. However, this picture represents fewer specialties in the research field being studied than that produced through the traditional first-author co-citation counting when the same number of top-ranked authors is selected and analyzed. Reasons for these effects are discussed
- …
