Truncated Newton method for sparse unconstrained optimization using automatic differentiation

When solving large complex optimization problems, the user is faced with three major problems. These are (i) the cost in human time in obtaining accurate expressions for the derivatives involved; (ii) the need to store second derivative information; and (iii), of lessening importance, the time taken to solve the problem on the computer. For many problems, a significant part of the latter can be attributed to solving Newton-like equations. In the algorithm described, the equations are solved using a conjugate direction method that only needs the Hessian at the current point when it is multiplied by a trial vector. In this paper, we present a method that finds this product using automatic differentiation while only requiring vector storage. The method takes advantage of any sparsity in the Hessian matrix and computes exact derivatives. It avoids the complexity of symbolic differentiation, the inaccuracy of numerical differentiation, the labor of finding analytic derivatives, and the need for matrix store. When far from a minimum, an accurate solution to the Newton equations is not justified, so an approximate solution is obtained by using a version of Dembo and Steihaug's truncated Newton algorithm (Ref. 1).This paper was presented at the SIAM National Meeting, Boston, Massachusetts, 1986.     

基于BDF的无约束优化方法的收敛性分析

1.介 绍 在上个世纪的七十年代末、八十年代初,基于常微分方程的优化方法或者说同伦方法是一类与拟牛顿法和共轭梯度法等我们所熟知的优化方法相竞争的重要方法[1-6,8,13,14,16].由于这类方法只是简单地利用现成的数值求解常微分方程的软件包,如CVODE[7]、LSODE[12],对同伦方程(一般是一个常微分方程的初值问题)进行计算,除了一些特殊的病态问题     

Sensitivity analysis of an optimized bar-spring model with hill-top branching

Summary Characteristics of optimal solutions under nonlinear buckling constraints are investigated by using a bar-spring model. It is demonstrated that optimization under buckling constraints of a symmetric system often leads to a structure with hill-top branching, where a limit point and bifurcation points coincide. A general formulation is derived for imperfection sensitivity of the critical load factor corresponding to a hill-top branching point. It is shown that the critical load is not imperfection-sensitive even for the case where an asymmetric bifurcation point exists at the limit point.     

An Application of Branch and Cut to Open Pit Mine Scheduling

The economic viability of the modern day mine is highly dependent upon careful planning and management. Declining trends in average ore grades, increasing mining costs and environmental considerations will ensure that this situation will remain in the foreseeable future. The operation and management of a large open pit mine having a life of several years is an enormous and complex task. Though a number of optimization techniques have been successfully applied to resolve some important problems, the problem of determining an optimal production schedule over the life of the deposit is still very much unresolved. In this paper we will critically examine the techniques that are being used in the mining industry for production scheduling indicating their limitations. In addition, we present a mixed integer linear programming model for the scheduling problems along with a Branch and Cut solution strategy. Computational results for practical sized problems are discussed.     

Determination of the portfolio selection for a property-liability insurance company

This paper presents an analysis of a portfolio model which can be used to assist a property-liability insurance company in determining the optimal composition of the insurance and investment portfolios. By introducing insurer's threshold risk and relaxing some non-realistic assumptions made in traditional chance constraint insurance and investment portfolio models, we propose a method for an insurer to maximize his return threshold for a given threshold risk level. This proposed model can be used to optimize the composition of underwriting and investment portfolios regarding the insurer's threshold risk level, as well as to generate the efficient frontier by adjusting insurer's threshold risk levels. A numerical example is given based on the industry's aggregated data for a sixteen year period.     

Magnetic properties of mechanically alloyed Co–Ti powder

An amorphous phase containing traces of non-transformed Co and Ti powders was obtained by mechanical alloying nominal compositions of Co₆₇Ti₃₃ and Co₅₀Ti₅₀ in a high-energy ball-mill. These alloys were prepared from elemental powders of Co and Ti. The heat treatment of Co₆₇Ti₃₃ at 573, 873 and 1173 K crystallized nanoparticles of Co₂Ti and Co₃Ti compounds, while the same treatments conducted on Co₅₀Ti₅₀ resulted in the formation of Co₂Ti and CoTi nanoparticles. The saturation magnetizations reached a maximum value in the amorphous state and they decreased when the temperatures of the heat treatment rose. Demagnetizing interparticle interaction effects were estimated through hysteresis loops and initial magnetization curves using the Fourier technique.     

Simplest optimal cutting patterns for equal rectangles

This paper presents branch-and-bound algorithms that can guarantee the simplest optimal cutting patterns of equal rectangles. An existing linear algorithm determines the global upper bound exactly. The branching process ends when a branch of a lower bound equal to the global upper bound is found.     

Structural manipulation of pyrochlores: Thermal evolution of metastable Gd2(Ti1−yZry)2O7 powders prepared by mechanical milling

The structural and microstructural characteristics of metastable Gd₂(Ti_1−yZr_y)₂O₇ powders prepared by mechanical milling have been studied by a combination of XRD and Raman spectroscopy. Irrespective of their Zr content, as-prepared powder phases present an anion-deficient fluorite-type of structure as opposed to the pyrochlore equilibrium configuration obtained for the same solid solution by other synthetic routes. These fluorites are stable versus thermal activation, at least up to temperatures of 800 °C. For the Ti-rich compositions, thermal treatments at higher temperatures facilitate the rearrangement of the cation and anion substructures and the relaxation of mechanochemically induced defects whereas for compositions with high Zr content, the fluorite crystal structure is retained even at temperatures as high as 1200 °C. Interestingly enough, transient pyrochlores showing a very unusual cation distribution were observed during the thermally induced defect-recovery process.     

Structural optimization design for large mirror

The system design must be started from the concept with low cost and high performance. In this point, the topological shape of the structure is very important in the view of the structural rigidities and light-weight design.In this paper, the optimization methodology is presented in the design stage of the large optical mirror. We obtain the optimal layout through the topology optimization and then design the details through the size or shape optimization for structural rigidity.     

C++QED: an object-oriented framework for wave-function simulations of cavity QED systems

We present a framework for efficiently performing Monte Carlo wave-function simulations in cavity QED with moving particles. It relies heavily on the object-oriented programming paradigm as realised in C++, and is extensible and applicable for simulating open interacting qua ntum dynamics in general. The user is provided with a number of “elements”, e.g. pumped moving particles, pumped lossy cavity modes, and various interactions to compose complex interacting systems, which contain several particles moving in electromagnetic fields of various configurations, and perform wave-function simulations on such systems. A number of tools are provided to facilitate the implementation of new elements.