|
|
Barkovsky E. A. (Petrozavodsk) Optimal
control of two queues working on the principle of custom queues 3
Abstracts . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Barkovsky
E. A. (Petrozavodsk) Optimal control of two queues
working on the principle of custom queues 33
|
||||
|
|
|
ABSTRACTS
Optimal Control of Two Queues Working on the Principle
of Custom Queues
E.
A. Barkovsky
Petrozavodsk
barkevgen@gmail.com
Key words: Data structures, custom queues, random walks, regular Markov chains.
In
this paper we propose a mathematical model and a solution of the problem of
optimal partitioning of shared memory for the custom queue in the form of two
queues (the case of their sequential cyclic representation). Operations of
insertion in one of the queues, with given probabilities, are performed on odd
steps, and operations of deletion are performed on even steps. It is possible
to execute operations both in serial and parallel ways.
Bibliogr.: 15 refs.
Invariants of Matrix Data in a Classification Problem
V. N. Shats
St.
Petersburg
vlnash@mail.ru
Key words: Artificial
intelligence, classification algorithms, data analysis, granular
computing.
This article presents a solution of a classification problem based on the calculation of index
matrices that are approximate invariants of the matrix data. Indices are the
numbers of intervals of feature values, and objects with the same index form
the granules. From the relation of the lengths of granules, we obtain the
frequency indices for features of any classes of the training sample which are
the same for the corresponding objects of the control sample. Then, for any
object, we evaluate the probability in each class and thereupon the most likely
object class. The high accuracy and stability of the results obtained using
this method are confirmed for nine data
sets from the UCI repository.
Bibliogr.: 13 refs.
