Software Journal:
Theory and Applications

Send article

Entrance Registration

Dear authors!

[09.12.2019]

Attention! The Journal has been renamed!

The former name is "International Scientific and Practical Journal "Software Products, Systems and Algorithms".

Due to implementing the Journal development plans, it will be published in English. The site content is undergoing certain changes.

We apologize for any inconvenience.
​Best Regards

All ads...

Solving the problem of placing the VLSI elements based on the integrating of swarm intelligence models into the affine search spaces

B.K. Lebedev (lebedev.b.k@gmail.com ) Institute of Computer Technology and Information Security of the Southern Federal University (Professor), Taganrog, Russian Federation, доктор технических наук;
O.B. Lebedev (lebedev.ob@mail.ru) Institute of Computer Technology and Information Security of the Southern Federal University (Associate Professor), Taganrog, Russian Federation, кандидат технических наук;
E.O. Lebedeva (lbedevakate@mail.ru ) Institute of Computer Technology and Information Security of the Southern Federal University (Postgraduate Student), Taganrog, Russian Federation;
A.A. Nagabedyan (andrewnagabedyan@yandex.ru ) Institute of Computer Technology and Information Security of the Southern Federal University (Graduate Student), Taganrog, Russian Federation;

The paper presents the architecture of a multi-agent system based on natural calculations, which places extra-large integrated circuits’ components using the combined swarm intelligence models. The authors offer new structures of presenting a solution for the problem of placing extra-large integrated circuit elements as chromosomes. There is a modified particle swarm paradigm that differs from the canonical one by the possibility of using the positions with integral-valued parameter values in the affine space.

A developed operator called directed mutation helps to move the swarm of particles in the observed solution area. The authors offer a modified structure of the bees algorithm. The key operation of the algorithm is the research on promising positions in the neighborhood of basic positions.

The tests have proven that when integrating the behavior models of a bee swarm and a particle swarm, the results of the new hybrid algorithm appear to be 11-18 % better than each algorithm results separately.