Time and Date: 11:00 - 12:40 on 12th June 2014
Room: Tully II
Chair: Leifur Leifsson
|94|| Fast Low-fidelity Wing Aerodynamics Model for Surrogate-Based Shape Optimization [abstract]
Abstract: Variable-fidelity optimization (VFO) can be efficient in terms of the computational cost when compared with traditional approaches, such as gradient-based methods with adjoint sensitivity information. In variable-fidelity methods, the direct optimization of the expensive high-fidelity model is replaced by iterative re-optimization of a physics-based surrogate model, which is constructed from a corrected low-fidelity model. The success of VFO is dependent on the reliability and accuracy of the low-fidelity model. In this paper, we present a way to develop a fast and reliable low-fidelity model suitable for aerodynamic shape of transonic wings. The low-fidelity model is component based and accounts for the zero-lift drag, induced drag, and wave drag. The induced drag can be calculated by a proper method, such lifting line theory or a panel method. The zero-lift drag and the wave drag can be calculated by two-dimensional flow model and strip theory. Sweep effects are accounted for by simple sweep theory. The approach is illustrated by a numerical example where the induced drag is calculated by a vortex lattice method, and the zero-lift drag and wave drag are calculated by MSES (a viscous-inviscid method). The low-fidelity model is roughly 320 times faster than a high-fidelity computational fluid dynamics models which solves the Reynolds-averaged Navier-Stokes equations and the Spalart-Allmaras turbulence model. The responses of the high- and low-fidelity models compare favorably and, most importantly, show the same trends with respect to changes in the operational conditions (Mach number, angle of attack) and the geometry (the airfoil shapes).
|Leifur Leifsson, Slawomir Koziel, Adrian Bekasiewicz|
|128|| Minimizing Inventory Costs for Capacity-Constrained Production using a Hybrid Simulation Model
Abstract: A hybrid simulation model is developed to determine the cost-minimizing target level for a single-item, single-stage production-inventory system. The model is based on a single discrete-event simulation of the unconstrained production system, from which an analytical approximation of the inventory shortfall is derived. Using this analytical expression it is then possible to evaluate inventory performance, and associated costs, at any target level. From these calculations, the cost-minimizing target level can be found efficiently using a local search. Computational experiments show the model remains highly accurate at high levels of demand variation, where existing analytical methods are known to be inaccurate. By deriving an expression for the shortfall distribution via simulation, no user modelling of the demand distribution or estimation of demand parameters is required. Thus this model can be applied to situations when the demand distribution does not have an identifiable analytical form.
|23|| Computation on GPU of Eigenvalues and Eigenvectors of a Large Number of Small Hermitian Matrices
Abstract: This paper presents an implementation on Graphics Processing Units of QR-Householder algorithm used to find all the eigenvalues and eigenvectors of many small hermitian matrices ( double precision) in a very short time to address time constraints for Radar issues.
|299|| COFADMM: A Computational features selection with Alternating Direction Method of Multipliers [abstract]
Abstract: Due to the explosion in size and complexity of Big Data, it is increasingly important to be able to solve problems with very large number of features. Classical feature selection procedures involves combinatorial optimization, with computational time increasing exponentially with the number of features. During the last decade, penalized regression has emerged as an attractive alternative for regularization and high dimensional feature selection problems. Alternating Direction Method of Multipliers (ADMM) optimization is suited for distributed convex optimization and distributed computing for big data. The purpose of this paper is to propose a broader algorithm COFADMM which combines the strength of convex penalized techniques in feature selection for big data and the power of the ADMM for optimization. We show that combining the ADMM algorithm with COFADMM can provide a path of solutions efficiently and quickly. COFADMM is easy to use, is available in C, Matlab upon request from the corresponding author.
|Mohammed Elanbari, Sidra Alam, Halima Bensmail|
|101|| Computational Optimization, Modelling and Simulation: Past, Present and Future [abstract]
Abstract: An integrated part of modern design practice in both engineering and industry is simulation and optimization. Significant challenges still exist, though huge progress has been made in the last few decades. This 5th workshop on Computational Optimization, Modelling and Simulation (COMS 2014) at ICCS 2014 will summarize the latest developments of optimization and modelling and their applications in science, engineering and industry. This paper reviews the past developments, the state-of-the-art present and the future trends, while highlighting some challenging issues in these areas. It can be expected that future research should focus on the data intensive applications, approximations for computationally expensive methods, combinatorial optimization, and large-scale applications.
|Xin-She Yang, Slawomir Koziel, Leifur Leifsson|