用次微分表达的对偶问题

本利用次微分建立了多目标规划的一个新的对偶问题,并给出其弱、强和逆对偶性,得到了一个新的次梯度的定义,并用其建立了一个新的对偶问题。     

Improved RGF method to find saddle points

The predictor-corrector method for following a reduced gradient (RGF) to determine saddle points [Quapp, W. et al., J Comput Chem 1998, 19, 1087] is further accelerated by a modification allowing an implied corrector step per predictor but almost without additional costs. The stability and robustness of the RGF method are improved, and the new version in addition reduces the number of gradient and Hessian calculations.     

How good is a Broyden–Fletcher–Goldfarb–Shanno-like update Hessian formula to locate transition structures? Specific reformulation of Broyden–Fletcher–Goldfarb–Shanno for optimizing saddle points

Based on a study of the Broyden–Fletcher–Goldfarb–Shanno (BFGS) update Hessian formula to locate minima on a hypersurface potential energy, we present an updated Hessian formula to locate and optimize saddle points of any order that in some sense preserves the initial structure of the BFGS update formula. The performance and efficiency of this new formula is compared with the previous updated Hessian formulae such as the Powell and MSP ones. We conclude that the proposed update is quite competitive but no more efficient than the normal updates normally used in any optimization of saddle points. © 1998 John Wiley & Sons, Inc. J Comput Chem 19: 349–362, 1998     

Position-space renormalization for systems with weak long-range interactions and the breakdown of hyperscaling

Exact renormalization group equations are derived for a position-space renormalization of spin systems with weak long-range forces. It is shown how an apparent dependence of the critical exponents on the choice of the renormalization group can be resolved via the mechanism of dangerous irrelevant variables and that this same mechanism is responsible for the breakdown of hyperscaling. The dimensiond=4 can be seen to be a borderline dimension above which classical critical exponents are expected.     

A modified saddle point method for predicting sound fields above a non-locally reacting porous medium

This paper examines an asymptotic analysis for predicting sound fields above a rigid-frame porous medium, the so-called non-locally reacting porous medium. Their solutions can be represented by a direct wave term, a reflected wave term and a diffraction wave term. Exact and analytical solutions are available for the direct wave and the reflected wave from a perfectly hard ground. In the contrary, the diffraction wave term can only be cast in an integral form that is amenable to approximate analysis. A modified saddle-point method is explored to evaluate the diffraction integral asymptotically. Three different types of non-locally reacting surfaces, which are an extended reaction, a hard-backed layer, and an impedance-backed layer, were considered. The sound fields above these porous surfaces have the same form but they are different by an augmented diffraction term in the solutions. The analytical formula for the total sound fields, which can be stated in a closed form, offer a physically interpretable solution comprising of a direct wave and ground reflected wave terms. This latter term can further be split into a specularly reflected plane wave and ground wave components. A series of numerical comparisons have been conducted to validate the asymptotic analyses presented in this study. It has been demonstrated that the overall sound fields can be predicted well by the formula for all incidence angles and for a wide range of non-locally reacting porous surfaces.     

Properties of Chaotic and Regular Boundary Crisis in Dissipative Driven Nonlinear Oscillators

The phenomenon of the chaotic boundary crisis and the related concept of the chaotic destroyer saddle has become recently a new problem in the studies of the destruction of chaotic attractors in nonlinear oscillators. As it is known, in the case of regular boundary crisis, the homoclinic bifurcation of the destroyer saddle defines the parameters of the annihilation of the chaotic attractor. In contrast, at the chaotic boundary crisis, the outset of the destroyer saddle which branches away from the chaotic attractor is tangled prior to the crisis. In our paper, the main point of interest is the problem of a relation, if any, between the homoclinic tangling of the destroyer saddle and the other properties of the system which may accompany the chaotic as well as the regular boundary crisis. In particular, the question if the phenomena of fractal basin boundary, indeterminate outcome, and a period of the destroyer saddle, are directly implied by the structure of the destroyer saddle invariant manifolds, is examined for some examples of the boundary crisis that occur in the mathematical models of the twin-well and the single-well potential nonlinear oscillators.     

Zero duality gap for a class of nonconvex optimization problems

By an equivalent transformation using thepth power of the objective function and the constraint, a saddle point can be generated for a general class of nonconvex optimization problems. Zero duality gap is thus guaranteed when the primal-dual method is applied to the constructed equivalent form.The author very much appreciates the comments from Prof. Douglas J. White.     

Complexity of local search for the p-median problem

We study the complexity of finding local minima for the p-median problem. The relationship between Swap local optima, 0–1 local saddle points, and classical Karush–Kuhn–Tucker conditions is presented. It is shown that the local search problems with some neighborhoods are tight PLS-complete. Moreover, the standard local descent algorithm takes exponential number of iterations in the worst case regardless of the tie-breaking and pivoting rules used. To illustrate this property, we present a family of instances where some local minima may be hard to find. Computational results with different pivoting rules for random and Euclidean test instances are discussed. These empirical results show that the standard local descent algorithm is polynomial in average for some pivoting rules.