Large-scale dynamical interconnections of stochastic singular systems (Q804526)

The paper deals with stochastic systems composed of a finite number of subsystems of the form \[ E_ i\dot x_ i=A_ ix_ i+B_ ia_ i+F_ is_ i,\quad b_ i=C_ ix_ i+D_ ia_ i+H_ iv_ i, \] where, relatively to the subsystem i, \(x_ i\) is the semistate, \(a_ i\) the input, \(b_ i\) the output, \(s_ i\) and \(v_ i\) the process and measurement noises, and \(E_ i\) is a square and generally singular matrix. Dynamic interconnections are allowed. This problem is more general than classical ones because of the joint presence of i) singular subsystems, ii) dynamic interconnections, iii) noise disturbances. The authors deal with the problem of well-definedness of the composite system (existence of a unique solution for all inputs), with new aspects in the stochastic case since derivatives of white noise may occur. Then they treat the question of finding a state variable representation of the composite system in the case of static interconnections. Some new structure algorithms for singular systems are used.