Wyniki wyszukiwania - BazTech

1

A Practical Approach to Trac Engineering using an Unsplittable Multicommodity Flow Problem with QoS Constraints

Białoń P.

Journal of Telecommunications and Information Technology

|

2016

|

nr 3

25--37

EN

The paper presents a practical approach to calculating intra-domain paths within a domain of a content-aware network (CAN) that uses source routing. This approach was used in the prototype CAN constructed as a part of the Future Internet Engineering project outcome. The calculated paths must satisfy demands for capacity (capacity for a single connection and for aggregate connections using the given path are considered distinctly) and for a number of path-additive measures like delay, loss ratio. We state a suitable variant of QoS-aware unsplittable multicommodity ow problem and present the solving algorithm. The algorithm answers to the needs of its immediate application in the constructed system: a quick return within a short and fairly predictable time, simplicity and modiability, good behavior in the absence of a feasible solution (returning approximately-feasible solutions, showing how to modify demands to retain feasibility). On the other hand, a certain level of overdimensioning of the network is explored, unlike in a typical optimization algorithm. The algorithm is a mixture of: (i) shortest path techniques, (ii) simplified reference-level multicriteria techniques and parametric analysis applied to aggregate the QoS criteria (iii) penalty and mutation techniques to handle the common constraints. Numerical experiments assessing various aspects of the algorithm behavior are given.

2

Solving Support Vector Machine with Many Examples

Białoń P.

Journal of Telecommunications and Information Technology

|

2010

|

nr 3

65-70

EN

Various methods of dealing with linear support vector machine (SVM) problems with a large number of examples are presented and compared. The author believes that some interesting conclusions from this critical analysis applies to many new optimization problems and indicates in which direction the science of optimization will branch in the future. This direction is driven by the automatic collection of large data to be analyzed, and is most visible in telecommunications. A stream SVM approach is proposed, in which the data substantially exceeds the available fast random access memory (RAM) due to a large number of examples. Formally, the use of RAM is constant in the number of examples (though usually it depends on the dimensionality of the examples space). It builds an inexact polynomial model of the problem. Another author's approach is exact. It also uses a constant amount of RAM but also auxiliary disk files, that can be long but are smartly accessed. This approach bases on the cutting plane method, similarly as Joachims' method (which, however, relies on early finishing the optimization).

3

A linear Support Vector Machine solver for a large number of training examples

Białoń P.

Control and Cybernetics

|

2009

|

Vol. 38, no 1

281-300

EN

A new linear Support Vector Machine algorithm and solver are presented. The algorithm is in a twofold way well-suited for problems with a large number of training examples. First, unlike many optimization algorithms, it does not simultaneously keep all the examples in RAM and thus does not exhaust the memory (moreover, it smartly passes through disk files storing the data: two mechanisms reduce the computation time by disregarding some input data without a loss in solution quality). Second, it uses the analytical center cutting plane scheme, appearing as more efficient for hard parameter settings than the Kelley's scheme used in other solvers, like SVM_perf. The experiments with both real-life and artificial examples are described. In one of them the solver proved to be capable of solving a problem with one billion training examples. A critical analysis of the complexity of SVM_perf is given.

4

Problems of Broadband in Rural Areas in Light of the BReATH Project Experiences

Białoń P.

Journal of Telecommunications and Information Technology

|

2008

|

nr 4

97-103

EN

Some lessons learned from the EU project "Broadband e-Services and Access to the Home" (2005-2007) are presented concerning the broadband development in rural ar- eas. In particular, the paper discusses the common problems of broadband deployment in the rural environment, various aspects of stimulating demand for broadband, the limitations of public aid and, most importantly, the problems of technoeconomic analysis.

5

An Eclectic Approach to Network Service Failure Detection Based on Multicriteria Analysis with an Example of Mixing Probabilistic Context Free Grammar Models

Białoń P.

Journal of Telecommunications and Information Technology

|

2008

|

nr 4

32-39

EN

A method of failure detection in telecommunication networks is presented. This is a meta-method that correlates alarms raised by failure-detection modules based on various philosophies. The correlation takes into account two main characteristics of each module and the whole metamethod: the percentage of false alarms and the percentage of omitted failures. The trade-off between them is tackled with aspiration-based multicriteria analysis. The alarms are correlated using linear classification by support vector machines. An example of the profitability of correlating alarms in such way is shown. This is an example of probabilistic context free grammars (PCFGs), used to model the proper runtime paths of network services (and thus usable for detecting an improper behavior of the services). It is shown that the linearly mixing PCFGs can add context handling to the PCFG model, thus augmenting the capabilities of the model.

6

Some variants of projection methods for large nonlinear optimization problems

Białoń P.

Journal of Telecommunications and Information Technology

|

2003

|

nr 3

43-49

EN

Two ideas of modifying projection methods for the case of smooth nonlinear optimization are presented. Projection methods were originally successfully used in solving large- scale linear feasibility problems. The proposed instantiations of projection methods fall into two groups. One of them is a decomposition approach in which projections onto sets are realized as optimization problems which themselves involve much portions of original problem constraints. There are two subproblems: one build with linear constraints of the original problem and the second one build with original nonlinear constraints. These approaches use special accelerating cuts so that the separation of nonlinear and linear constraints can be effective and some problem sparsity preserved. The second group bases on penalty-shifting/multiplier methods and draws from the observation that unconstrained subproblems obtained there may solve very slowly due to their nonsmooth character. Thus it is proposed to solve them with modified projection methods which inherit from conjugate gradient methods a multi-dimensional subspace in one epoche.