The Sequential Minimal optimization (SMO) algorithm is a popular algorithm used to solve the Support Vector Machine problem due to its efficiency and ease of implementation. We investigate applying extrapolation metho...
详细信息
ISBN:
(纸本)0780382439
The Sequential Minimal optimization (SMO) algorithm is a popular algorithm used to solve the Support Vector Machine problem due to its efficiency and ease of implementation. We investigate applying extrapolation methods to the SMO update method in order to increase the rate of convergence of this algorithm. We first show that the update method is Newtonian and that extrapolation ensures the update is norm reducing on the objective function. We also note that choosing the working set pair according to some partial order does result in slightly faster speedups in algorithm performance.
The effect of a repair of a complex system can usually be approximated by the following two types: minimal repair for which the system is restored to its functioning state with minimum effort, or perfect repair for wh...
详细信息
The effect of a repair of a complex system can usually be approximated by the following two types: minimal repair for which the system is restored to its functioning state with minimum effort, or perfect repair for which the system is replaced or repaired to a good-as-new state. When both types of repair are possible an important problem is to determine the repair policy;that is, the type of repair which should be carried out after a failure. In this paper, an optimal. Allocation problem is studied for a monotonic failure rate repairable system under some resource constraints. In the first model, the numbers of minimal & perfect repairs are fixed, and the optimal repair policy maximizing the expected system lifetime is studied. In the second model, the total amount of repair resource is fixed and the costs of each minimal & perfect repair are assumed to be known. The optimal allocation algorithm is derived in this case. Two numerical examples are shown to illustrate the procedures.
An optimal algorithm for the one-stage estimation of the radiator position by a passive system consisting of narrow-base subsystems was developed. A comparative analysis of one-stage and direction-finding (two-stage) ...
详细信息
An optimal algorithm for the one-stage estimation of the radiator position by a passive system consisting of narrow-base subsystems was developed. A comparative analysis of one-stage and direction-finding (two-stage) methods was carried out using both the Cramer-Rao matrix bound and computer simulation. It was shown that one-stage and direction-finding methods offer equal accuracies of measuring the radiator coordinates if the signal-to-noise ratio exceeds some threshold value. It was revealed that, for low values of the signal-to-noise ratio, the one-stage method is far more resistant to abnormal errors than the direction-finding method.
An Advanced Traveler Information System (ATIS) stand-alone operation scheme is formulated as a bi-level optimization problem. The scheme logic attempts to optimize the network overall travel time by adjusting the path...
详细信息
An Advanced Traveler Information System (ATIS) stand-alone operation scheme is formulated as a bi-level optimization problem. The scheme logic attempts to optimize the network overall travel time by adjusting the path proportions while guessing the signal phase split decisions. An approximate simulation-based optimization algorithm is devised as an example of the logic operating this scheme. The logic is then replicated by a fuzzy-logic control system. Neural nets are utilized to develop the knowledge base of the fuzzy system and to calibrate the fuzzy set parameters. The neural nets utilize data replicates generated by the approximate simulation-based optimization algorithm. The calibration and effectiveness results of the fuzzy control system are presented.
In addition to texture, plastic anisotropy of a polycrystalline fcc metal stems from the directional nature of the dislocation substructure within individual grains. This produces the marked work hardening or softenin...
详细信息
In addition to texture, plastic anisotropy of a polycrystalline fcc metal stems from the directional nature of the dislocation substructure within individual grains. This produces the marked work hardening or softening observed immediately following load path changes. Following the framework of Peeters et al.,([1.2]) in bcc steel, we develop a dislocation substructure evolution-based stage III hardening model for copper, capable of capturing the constitutive response under load path changes. The present model accounts for the more complicated substructure geometry in fcc metals than in bee. Using an optimization algorithm, the parameters governing substructure evolution in the model are fit to experimental stress-strain curves obtained during compression along the three orthogonal directions in samples previously rolled to various reductions. These experiments approximate monotonic, reverse, and cross-load paths. With parameters suitably chosen, the substructure model, embedded into a self-consistent polycrystal plasticity model, is able to reproduce the measured flow stress response of copper during load path change experiments. The sensitivity of the parameters to the assumed substructure geometry and their uniqueness are also discussed.
The presence of a space-invariant sinusoidal phase error in the azimuth SAR (synthetic aperture radar) signal history causes paired echoes to appear in the system impulse response [1]. These high peak sidelobes may be...
详细信息
The presence of a space-invariant sinusoidal phase error in the azimuth SAR (synthetic aperture radar) signal history causes paired echoes to appear in the system impulse response [1]. These high peak sidelobes may be interpreted erroneously as separate targets by conventional autofocus algorithms. In this paper, a new approach to solve the high-frequency phase noise problems in SAR imaging is presented. The phase error is formulated as a nonlinear optimization problem for continuous variables. The optimum solution is obtained by minimizing the entropy of the signal. The proposed algorithm utilizes a real-coded genetic algorithm (GA) in the first stage to direct the search towards the global optimum region and a local search method in the second stage to do fine tuning. The algorithm is tested on simulated point targets and the results show significant improvement in the target response, as compared to the conventional autofocus and optimization algorithms. Thus the efficiency and the success rate of the algorithm are placed in sharper focus to minimize high frequency phase noise. Furthermore, the complexity of the problems solved clearly demonstrates the usefulness of this novel approach in solving such nonlinear systems.
We review the space-mapping (SM) technique and the SM-based surrogate (modeling) concept and their applications in engineering design optimization. For the first time, we present a mathematical motivation and place SM...
详细信息
We review the space-mapping (SM) technique and the SM-based surrogate (modeling) concept and their applications in engineering design optimization. For the first time, we present a mathematical motivation and place SM into the context of classical optimization. The aim of SM is to achieve a satisfactory solution with a minimal number of computationally expensive "fine" model evaluations. SM procedures iteratively update and optimize surrogates based on a fast physically based "coarse" model. Proposed approaches to SM-based optimization include the original algorithm, the Broyden-based aggressive SM algorithm, various trust-region approaches, neural SM, and implicit SM. Parameter extraction is an essential SM subproblem. It is used to align the surrogate (enhanced coarse model) with the fine model. Different approaches to enhance uniqueness are suggested, including the recent gradient parameter-extraction approach. Novel physical illustrations are presented, including the cheese-cutting and wedge-cutting problems. Significant practical applications are reviewed.
Formal treatment of collaborative multi-agent systems has been lagging behind the rapid progress in sequential decision making by individual agents. Recent work in the area of decentralized Markov Decision Processes (...
详细信息
Formal treatment of collaborative multi-agent systems has been lagging behind the rapid progress in sequential decision making by individual agents. Recent work in the area of decentralized Markov Decision Processes (MDPs) has contributed to closing this gap, but the computational complexity of these models remains a serious obstacle. To overcome this complexity barrier, we identify a specific class of decentralized MDPs in which the agents' transitions are independent. The class consists of independent collaborating agents that are tied together through a structured global reward function that depends on all of their histories of states and actions. We present a novel algorithm for solving this class of problems and examine its properties, both as an optimal algorithm and as an anytime algorithm. To the best of our knowledge, this is the first algorithm to optimally solve a non-trivial subclass of decentralized MDPs. It lays the foundation for further work in this area on both exact and approximate algorithms.
There are two basic types of interconnection agreements between providers in the Internet: peering and transit. A decision every Internet network service provider (INSP) has to make is which other peering/transit INSP...
详细信息
There are two basic types of interconnection agreements between providers in the Internet: peering and transit. A decision every Internet network service provider (INSP) has to make is which other peering/transit INSPs to connect with. The potential peering/transit partners offer (obviously) different routes and they may differ quite drastically in the amount and type of charges (line costs, exchange point related costs, settlement costs, administrative costs) they demand as well as in reliability and quality of service. In this article, we discuss and solve problems in this context: the first problem is finding the optimal set of peering and transit partners for one INSP at one point in time, given the routing information and the cost functions of the potential peering/transit partners and the Internet exchange points. Reliability issues (for example enforcing enough spare capacity to absorb the complete failure of one provider) are considered as well as quality of service constraints (e.g., enforcing a certain average AS-hop count). These extended problems are formally described and solved with an optimal algorithm and compared with heuristics. Rates and traffic are in a permanent state of flux, thus an INSP always has to consider whether its current choice of peering/transit partners is still optimal or whether it is worth changing some of its peering/transit agreements. The last part of this article deals with this problem and adapts the algorithm from the first part for this setting. (C) 2004 Elsevier B.V. All rights reserved.
A new approach based on the decomposition, in position-time space, of operating functions (OFs) is proposed and various criteria of parameter identification like the output least squares (OLS) and the regularized leas...
详细信息
A new approach based on the decomposition, in position-time space, of operating functions (OFs) is proposed and various criteria of parameter identification like the output least squares (OLS) and the regularized least squares (RLS) are formulated and incorporated in the framework. Several optimization algorithms such as the Gauss-Newton, conjugate gradient and Levenberg-Marquardt are diligently modified and computed in order to search the performed algorithm to identify, in space and time, the numerical fields of the model parameters. Because the sensitivity of the criteria to the design variables plays a very important role in the optimization problem, various techniques such as finite difference method and direct differentiation method are tested. The enhanced performance of the model with the newly identified OFs is proved by a higher conformity of its predictions with the real data of drying wood systems (DWSs).
暂无评论