Vector Stochastic Differential Equations Used to

Electrical Networks with Random Parameters

Edita Kolářová and Lubomı́r Brančı́k

Abstract—In this paper we present an application of the Itô II. S TOCHASTIC D IFFERENTIAL E QUATIONS
stochastic calculus to the problem of modelling RLC electrical
A. Vector Stochastic Differential Equations
circuits. The deterministic model of the circuit is replaced by a
stochastic model by adding a noise term to various parameters In the theory of stochastic differential equations the Wiener
of the circuit. The analytic solutions of the resulting stochastic process plays a very important role, because it represents the
integral equations are found using the multidimensional Itô integral of the so called Gaussian white noise, that describes
formula. For the numerical simulations in the examples we
used MATLABr . The SDE approach has its perspectives in the the randomness in stochastic models of physical events.
simulation even higher-order circuits representing more complex A real-valued Wiener process W (t) is a continuous stochas-
physical systems, as their real implementations are often subject tic process with independent increments, W (0) = 0 and
to a number of random effects. An example for a transmission W (t) − W (s) distributed N (0, t − s), 0 ≤ s < t. Notice
line lumped-parameter model is provided. in particular that E[W (t)] = 0, E[W 2 (t)] = t, t ≥ 0.
Keywords—Stochastic differential equations, Wiener process, We can define an N-dimensional SDE in vector form as
Itô formula, electrical network, transmission line model. XM
dX(t) = A(t,
e X(t)) dt + fj (t, X(t)) dW j (t), (1)
B
I. I NTRODUCTION j=1

Stochastic differential equations (SDEs) describe systems where Ae : h0, T i × RN → RN is a vector function, B fj rep-
including some random effects. In this paper we deal with resents the j−th column of the matrix function B : h0, T i ×
f
vector Itô stochastic differential equations and refer to some RN → RN ×M and dW(t) = ( dW 1 (t), . . . , dW M (t)) is
utilization in electrical engineering simulations. Using the a column vector, where W 1 (t), . . . , W M (t) are independent
Itô formula we find the analytic solution of the equations Wiener processes representing the noise. The solution is a
containing one stochastic parameter. Then we apply the theory stochastic vector process X(t) = (X 1 (t), . . . , X N (t)). By
to the stochastic model of an RLC electrical network that we an SDE we understand in fact an integral equation
get by replacing one of its parameters in the deterministic
Z t M Z t
model by a random process. The theory of SDEs can find X fj (s, X(s)) dW j (s),
X(t) = X0 + e X(s)) ds+
A(s, B
its place in various fields of the science and engineering, t0 t0
j=1
when random effects are to be considered [1]-[4]. In field of (2)
the electrical engineering it can cover a number of random where the integral with respect to ds is the Lebesgue integral
processes occurring in electrical systems, see e.g. [5]-[7] and the integrals with respect to dW j (s) are stochastic
to mention at least a few application areas. Some attention integrals, called the Itô integrals (see [2]).
is still paid to the first-order RL or RC circuits, see e.g. Although the Itô integral has some very convenient proper-
[8]-[11], where suitable numerical techniques and different ties, the usual chain rule of classical calculus doesn’t hold. The
noise types have been studied. From practical point of view, appropriate stochastic chain rule is known as the Itô formula.
however, the second-order RLC or RLGC circuits are of major
importance as they serve as building blocks of more complex B. The Multidimensional Itô Formula
physical models. In this paper we want to prepare a bases
Let the stochastic process X(t) be a solution of the vector
for the solution of interconnects under stochastically varied
stochastic differential equation (1) for some suitable functions
parameters, modelled by a cascade connection of just the 2nd-
A, B (see [2], p.48). Let g(t, X) : (0, ∞) × RN → RP be a
order networks [12], [13] and do first verifications. Whilst for
twice continuously differentiable function. Then
the 1st-order models a scalar SDEs theory is commonly used,
see [9], in case of higher-order models the vector SDEs are Y(t) = g(t, X(t)) = (g1 (t, X), . . . , gP (t, X)) (3)
needed to be applied. is a stochastic process, whose k−th component is given by
Manuscript received October 28, 2012. This work was supported by the ∂gk X ∂gk
project no. CZ.1.07/2.3.00/20.0007 WICOMT of the operational program dY k = (t, X) dt + (t, X) dX i + (4)
Education for competitiveness, and was performed in laboratories supported ∂t i
∂xi
by the SIX project, no. CZ.1.05/2.1.00/03.0072, the operational program
Research and Development for Innovation. 1 X ∂ 2 gk
Edita Kolářová is with Brno University of Technology, Department of Math- + (t, X)( dX i )( dX j ),
2 i,j ∂xi ∂xj
ematics, Technicka 8, Brno, Czech Republic. (e-mail: [email protected]).
Lubomı́r Brančı́k is with Brno University of Technology, Department of
Radio Electronics, Purkynova 118, Brno, Czech Republic. (e-mail: bran- where dX i · dX j is computed according to the rules dt· dt =
[email protected]). dt · dW i = dW i · dt = 0 and dW i · dW j = δi,j dt.
C. Vector Linear Itô Stochastic Differential Equations E. Solution with Multiplicative Noise
The stochastic differential equation (1) is called linear, We will solve the linear stochastic equation :
provided the coefficients have the form

e X(t)) = A(t)X(t) + a(t), dX(t) = A X(t) + a(t) dt + B X(t) dW (t), (12)
A(t, (5)
and where A and B are 2×2 matrices and a(t) : h0, T i → R2 is a
vector function.
fj (t, X(t)) = Bj (t) X(t) + bj (t),
B j = 1...M (6) We define a function g(t, x1 , x2 , y) : h0, T i × R3 → R2 by
for A, Bj : h0, T i → RN ×N and a, bj : h0, T i → RN .
 
1 2 x1
For the applications to the RLC(G) circuit we need only a g(t, x1 , x2 , y) = e( 2 B −A)t−By (13)
x2
2 dimensional vector linear equation with a one dimensional
Wiener process, which has the form and compute the derivative of g(t, X(t), W (t)) by the Itô


formula.
dX(t) = A X(t) + a(t) dt + B X(t) + b(t) dW (t), (7)  1 2 
d e( 2 B −A)t−BW (t) X(t) =
where A and B are 2×2 matrices, a : h0, T i → R2 and
similarly b : h0, T i → R2 are vector functions, W (t) is the 
1 2

( 21 B2 −A)t−BW (t)
Wiener process. =e B − A X(t) dt+
2
1 2
D. Solution with Additive Noise −A)t−BW (t)
+e( 2 B (−B)X(t) dW (t)+
First we solve the equation (7) with additive noise, which 2
1
−A)t−BW (t)
is the case when B ≡ 0. To find the analytic solution of the +e( 2 B dX(t)+
equation
1 1 2
+ e( 2 B −A)t−BW (t) (−B)(−B) d2 W (t)+


dX(t) = A X(t) + a(t) dt + b(t) dW (t), (8) 2
we use the multidimensional Itô formula and find the derivative 1 1 2
+ e( 2 B −A)t−BW (t) (−B) dX(t) · dW (t) +


of the function g(t, X(t)) = e−At X(t) : h0, T i × R2 → R2 :
2
dg(t, X(t)) = d e−A t · X(t) =


1 1 2
+ e( 2 B −A)t−BW (t) (−B) dW (t) · dX(t) =


= −Ae−A t · X(t) dt + e−A t · dX(t) − 0 · ( dX(t))2 = 2
= −Ae−A t X(t) dt + e−A t AX(t) dt+ 1 2
1
= e( 2 B −A)t−BW (t) B2 X(t) dt−AX(t) dt−BX(t) dW (t)+
2
+ e−At a(t) dt + e−At b(t) dW (t).
1
We have +AX(t) dt + a(t) dt + BX(t) dW (t) + B2 X(t) dt+
2
d e−A t · X(t) = e−At (a(t) dt + b(t) dW (t)) .


1 1 
Integrating the last equation we get + (−B)BX(t) dt + (−B)BX(t) dt =
2 2
Z t Z t
1 2
e−A t ·X(t)−X(0) = e−As a(s) ds+ e−As b(s) dW (s). = e( 2 B −A)t−BW (t)
a(t) dt.
0 0
From this we can easily get the solution We get
Z t Z t  1 2  1 2
( B −A)t−BW (t)
X(t) = eAt X(0)+ eA(t−s) a(s) ds+ eA(t−s) b(s) dW (s). d e 2 X(t) = e( 2 B −A)t−BW (t) a(t) dt. (14)
0 0
(9) After the integration and some computation as in section D,
The solution X(t) is a random process and for its expectation we get the solution
we have for every t > 0
1 2
Z t X(t) = e(A− 2 B )t+BW (t)
X(0)+
E[X(t)] = eA t · E[X(0)] + eA(t−s) a(s) ds, (10)
0
t
while the expectation of the Itô integral is zero. We can
Z
1 2

see, that for constant initial value X(0), which is usually e(A− 2 B )(t−s)+B(W (t)−W (s))
a(s) ds. (15)
0
the case, the expectation of the stochastic solution coincides
with the deterministic solution of the equation (8). By the If X(0) is constant, the expectation of the solution, E[X(t)]
deterministic solution we mean the analytic solution of the for t ∈ h0, T i is the unique solution of the ordinary differential
ordinary differential equation equation, (see [1])



dX(t) = A X(t) + a(t) dt. (11) dE[X(t)] = A E[X(t)] + a(t) dt. (16)
 III. RLC E LECTRIC C IRCUIT C. RLC Circuit with Stochastic Resistance
A. Deterministic Model We consider now the resistance influenced by random
Let Q(t) be the charge at time t at a fixed point in an electric effects. Instead of R we have:
circuit, and let L be the inductance, R the resistance and U (t)
R∗ = R + “noise” = R + αξ(t), (24)
the potential source at time t. The charge Q(t) satisfies the
differential equation where ξ(t) denotes the “white noise process”, α is a constant.
1 We put this to the form(17) and rewrite the second order
LQ00 (t) + RQ0 (t) + Q(t) = U (t), (17) equation as a system of two equations, then multiply both
C
with initial conditions Q(0) = Q0 , Q0 (0) = I0 . We can equations by dt and replace ξ(t) dt by dW (t) . We get the
transform this equation by introducing the vector stochastic vector equation (12)


dX(t) = A X(t) + a(t) dt + B X(t) dW (t),
   
X1 (t) Q(t)
X(t) = =
X2 (t) Q0 (t)  
0 0
to the system where B = α , A and a(t) are as in (19), with the
0 −L
solution (15).
X10 = X2
The expectation of X(t) solves the equation (19), so the
1 R U (t)
X20 = − X1 − X2 + . (18) expectation E[X(t)] is equal to the deterministic solution (20).
CL L L
This in matrix notation gives
D. SDE Numerical Methods Applied
dX(t)
= A · X(t) + a(t), X(0) = X0 (19) For practical use of stochastic differential equations we have
dt
to simulate the stochastic solution by numerical techniques.
where
      The simplest numerical scheme, the stochastic Euler scheme,
0 1 0 Q0 is based on the Euler numerical scheme for ordinary differen-
A= 1 , a(t) = U (t) , X0 = .
− LC −RL L
I0 tial equations.
The equation (19) has the analytic solution Let us consider an equidistant discretisation of the time
Z t interval h0, T i, namely
X(t) = eAt X0 + eA(t−s) a(s) ds. (20) Z tk+1
0 T
tk = kh, where h = = tk+1 − tk = dt,
K tk
B. RLC Circuit with Stochastic Source
K ∈ N, k = 0, . . . , K − 1 and the corresponding discretisa-
We will consider the source influenced by random effects.
tion of the j−th component of the Wiener process,
Instead of U (t) we consider the non deterministic version of
Z tk+1
this function:
∆Wkj = W j (tk+1 ) − W j (tk ) = dW j (s).
U ∗ (t) = U (t) + “noise”. (21) tk

To be able to substitute this into the equation (17) we have To be able to apply a stochastic numerical scheme, first we
to describe mathematically the “noise”. It is reasonable to have to generate, for all j, the random increments of W j as
look at it as a stochastic process called the “white noise independent Gauss random variables with mean E[∆Wkj ] = 0
process”, denoted by ξ(t). We get the following equation (α and E[(∆Wkj )2 ] = h.
is a constant) The explicite Euler scheme for the i−th component of
  the N dimensional stochastic differential equation (1) has the
dX(t) 0
= A · X(t) + a(t) + α . (22) form
dt L ξ(t) M
X
We multiply (22) by dt and then replace ξ(t) dt by dW (t); i
Xk+1 = Xki + Ai (tk , Xk )h + B i,j (tk , Xk )∆Wkj . (25)
W (t) is the Wiener process. Formally the “white noise” is the j=1
time derivative of the Wiener process W (t). We get
  For measuring the accuracy of a numerical solution to an SDE
  0 we use the strong order of convergence. The Euler scheme
dX(t) = A · X(t) + a(t) dt + α .
L dW (t) converges with strong order 21 (see [3]).
We got the equation of the form(8)
 
  0
dX(t) = A·X(t)+a(t) dt+b dW (t), where b = α ,
L
with the solution
Z t Z t
X(t) = eAt X(0) + eA(t−s) a(s) ds + b eA(t−s) dW (s).
0 0
(23)
and the expectation equal (20). Fig. 1. RLC(G) circuit with unit-step voltage source.
 IV. E XAMPLES
A. Example 1
First we consider the RLC circuit, see Fig. 1, excited from
a unit-step voltage source influenced by a noise,
1∗ (t) = 1(t) + αξ(t),
with remaining circuit parameters deterministic.
The values of all the parameters are unit which corresponds
to an underdamped behavior of the circuit. Hereafter we con-
sider the capacitor voltage VC (t) = Q(t)/C and the inductor
current IL (t) = Q0 (t) as state variables in the circuit instead of
the charge-based notations in (17)-(19), which will be useful
for further considerations. Also zero initial conditions are
considered in this example.
Current and voltage responses, namely their 100 realiza-
tions, including their sample means accompanied by 99%
confidence intervals highlighted, are presented in Fig. 2.
Deterministic solutions based on (20) lead to formulae
1 −βt
iL (t) = e sin ωt (26)
ωL
 
β
vC (t) = 1 − e−βt sin ωt + cos ωt (27)
ω
with r
R 1
β= and ω = − β2.
2L LC
Fig. 3. RLC circuit stochastic responses (α = 0.15, noisy resistor).

It can easily be confirmed correspondences between the sam-

ple means in Fig. 2 and waveforms based on (26) and (27).
Some dissimilarities in the above stochastic processes can
be observed, ”smoother” character of the voltages is likely
caused by a filtering effect of the low-pass filter formed by
this network.

B. Example 2
The second example considers the RLC circuit excited from
a deterministic unit-step voltage source 1(t), with L and C
deterministic as well, but with R influenced by a noise, i.e.
its resistance is equal R∗ (t) = R + αξ(t).
The current and voltage responses (their 100 realizations),
including their sample means with 99% confidence intervals
highlighted, are depicted in Fig. 3. The voltage stochastic
waveforms look again ”smoother” than the current ones, even
more when comparing with Fig. 2, and with less dispersions
for the same noise intensity factor α.

C. Example 3
The third example considers an RLCG circuit, or the RLC
circuit terminated by a resistive load . The conductance G can
also model nonideal properties of a real electrical condenser,
namely its nonzero leakage. In this case the matrix A defined
in (19) changes as
!
G
−C 1
A= 1
,
Fig. 2. RLC circuit stochastic responses (α = 0.15, noisy source).
− LC −RL
 C0 l/m, and Gd = G0 l/m, where m is a number of Π
sections in cascade. For the Thévenin resistances of terminal
circuits supposed as nonzero, the terminal currents are given
by iS = (vS − v1 )/RS , and iL = vL /RL = vm+1 /RL . The
asterisks at an exciting voltage, vS∗ , and series resistances, Rd∗ ,
mean that respective quantity is considered as stochastically
varying.

A. Deterministic Model
As is shown e.g. in [16] a uniform TL m-sectional model
can be described by a state-variable method leading to a vector
ordinary differential equation
dx(t)
= Ax(t) + Bu(t) (28)
dt
with a well-known analytical solution
Z t
x(t) = eAt x0 + eA(t−s) Bu(s) ds (29)
0
Formal similarities with (19) and (20) suggest us to proceed
with stochastic solutions by similar way, see next chapters.
The individual terms in (28) will be formulated as follows.
The matrices A = −M−1 (H + P), and B = M−1 P, where
! ! !
C 0 G E YE 0
M= , H= , P= , (30)
0 L -ET R 0 0
with C = Cd Im+1 , L = Ld Im , G = Gd Im+1 , and R = Rd Im ,
as diagonal matrices (there are exceptions in C and G related
to boundary elements, when Cd /2 and Gd /2 are used), I
Fig. 4. RLCG circuit stochastic responses (α = 0.15, noisy resistor).
identity matrices of indexed orders, and 0 corresponding zero
matrix. The E is a special (m + 1) × m bidiagonal matrix
remaining matrices keep their forms. When a conductance is containing ±1 and zeros, see [16] for more details, and
chosen e.g. as G = 0.2S, the circuit stays underdamped, with the (m + 1)-order matrix YE = diag(GS , 0, ..., 0, GL ) is
analytical solutions presented e.g. in [14]. Numerical solutions dependent on the external circuits, where GS = RS−1 and
−1
for stochastic exciting voltage source are, from qualitative GL = R L . The column vector x(t) = [vTC (t), iTL (t)]T consists
viewpoint, very similar to those in Fig. 2. In case of stochastic of the state variables required, namely of subvectors vC (t) and
resistance R, however, the results are presented in Fig. 4. Here iL (t) holding m+1 capacitor voltages and m inductor currents,
it is obvious that due to a permanent current flowing through respectively. Finally, the column vector u(t) = [vTE (t), 0]T
the resistor R, stochastic responses do not tend to be damped contains the (m + 1)-order vector vE (t) = [vS (t), 0, ..., 0]T
down to zero, unlike the RLC circuit in Fig. 3. acting as an excitation term.

V. U NIFORM T RANSMISSION L INE M ODELLING B. Model with Stochastic Excitation

Here a uniform transmission line (TL) lumped-parameter Here, instead of vS (t), a non-deterministic version is used
model will be considered and its responses simulated, see
Fig. 5. If we denote l as a TL length and L0 , R0 , C0 and vS∗ (t) = vS (t) + αξ(t). (31)
G0 as its per-unit-length parameters the lumped parameters of
while remaining parameters are deterministic.
the model are defined by Ld = L0 l/m, Rd = R0 l/m, Cd =

Fig. 5. Uniform transmission line m-sectional lumped-parameter model.

A noise term ξ(t) is a stochastic process, namely a white D. SDE Numerical Method Applied
noise, and with α a constant, expressing its intensity. After As computer simulations revealed, the explicit Euler scheme
substitution (31) into (28) and doing arrangements keeping (25) satisfies only for relatively low orders of electric models
the SDE theory above we get a vector linear SDE with an under consideration to get stable solutions. That is why,
additive noise implicit Euler schemes will be used for all the practical simula-
tions of the TL model.Then, based on [3] and considering (32)
dX(t) = (AX(t) + Bu(t)) dt + b dW (t) (32) and (36), implicit stochastic Euler schemes can be formulated
respectively as
with b = B[αT T
E , 0] , and αE = [α, 0, . . . , 0]
T
as a noise −1
Xn+1 = (I − Ah) Xn + Bun+1 h + b∆W n


intensity vector of the order m + 1. The stochastic solution (38)
X(t) in (32) is marked by a capital letter to distinguish it for the SDE with an additive noise, and
−1
Xn+1 = (I − Ah) Xn + Bun+1 h + DnW Xn


from that deterministic in (29). After some modification of the (39)
theory from section two, considering the multidimensional Itô for the SDE with a multiplicative noise, where I is the identity
formula [1], [2], the formal analytical stochastic solution is matrix, being consistent with the Itô stochastic calculus.
Z t Z t
X(t) = eAt X0 + eA(t−s) Bu(s) ds + eA(t−s) b dW (s), E. Examples of TL Model Responses
0 0
(33) Let us now consider a uniform TL with per-unit-length
with the right term called as the Itô integral. The X(t) is a parameters R0 = 0.1Ω/m, L0 = 494.6nH/m,G0 = 0.1S/m,
random process and for its expectation we have for t > 0 and C0 = 62.8pF/m [15]. The TL’s length l = 0.3m, the
terminal resistances RS = 10Ω and RL = 1kΩ, the drive
t
source is a sine-squared impuls, vS (t) = sin2 (πt/2 · 10−9 ) if
Z
At
E[X(t)] = e E[X0 ] + eA(t−s) Bu(s) ds, (34)
0
0 ≤ t ≤ 2 · 10−9 s, and vS (t) = 0 otherwise. To demonstrate
overall state on the TL the deterministic solutions for voltage
when the expectation of the Itô integral is zero. It is evident and current distributions are depicted in Fig. 6, computed
that for an initial value X0 being constant, this expectation through a continuous model in a Laplace domain to be more
coincides with the deterministic solution (29). accurate for a comparison, and by using a numerical inversion
of Laplace transforms [15].

C. Model with Stochastic Resistances

Herein the responses to a deterministic excitation vS (t) but
under all m resistances Rd in the TL model influenced by
random effects are considered. Instead of the original Rd , their
non-deterministic versions are used as
∗
Rdk (t) = Rd + αk ξk (t). (35)

k = 1, . . . , m, where noise terms ξk (t) are again considered as

”white noise processes”, and αk as their intensities. Notice that
the series resistances connected to different nodes are generally
affected by different noise terms like it can occur on a real TL.
After substitution (35) into (28), and keeping the SDE theory
procedures, we get

dX(t) = (AX(t) + Bu(t)) dt + DW (t)X(t) (36)

with !
−1 0 0
DW = −M , (37)
0 αR dWR (t)

where the matrices αR = diag(α1 , ..., αm ) and dWR (t) =

diag( dW1 (t), ..., dWm (t)). This is a general vector linear
SDE with a multiplicative noise, which is represented by a m-
dimensional Wiener process. The solution in general doesn’t
have the normal distribution for X0 constant see [1], but the
expectation of the solution (36) is given again by (34), its exact
analytical solution can be obtained by generalizing results
shown e.g. in [17]. Fig. 6. Voltage and current distributions along the uniform TL.
Fig. 7. Stochastic voltage responses for noisy excitation voltage. Fig. 9. Stochastic voltage responses for noisy series resistances.

Fig. 8. Stochastic current responses for noisy excitation voltage. Fig. 10. Stochastic current responses for noisy series resistances.
 The Fig. 7 and 8 show a case of additive noises, Fig. 9 R EFERENCES
and 10 then a case of multiplicative ones. The computation [1] L. Arnold, Stochastic Differential Equations: Theory and Applications,
was performed for m = 128 sections, which complies, with New York: John Wiley & Sons, 1974.
some margin, with rule of thumb the TL section delay is [2] B. Øksendal, Stochastic Differential Equations, An Introduction with
Applications, New York: Springer-Verlag, 2000.
less then roughly a tenth of the rise/fall time of the impuls [3] P. Kloeden, E. Platen, and H. Schurz, Numerical Solution of SDE
propagated (m > 100 would be good in our case, a choice Through Computer Experiments, New York: Springer-Verlag, 1997.
128 simplifies the usage of another comparative method, a 2D [4] S. Cyganowski, P. Kloeden, and J. Ombach, From Elementary Probabil-
ity to Stochastic Differential Equations with Maple, New York: Springer-
LT in conjunction with an FFT-based 2D-NILT algorithm). Verlag, 2000.
The 99% confidence intervals (dashed red lines), marking [5] D. Ham, and A. Hajimiri, ”Complete Noise Analysis for CMOS Switch-
uncertainities in determining the sample mean values (solid ing Mixers via Stochastic Differential Equations,” in Proc. IEEE 2000
Custom Integrated Circuits Conference, Orlando, pp. 439-442., 2000.
blue lines) are highlighted, together with the confidence in- [6] K. Wang, and M. L. Crow, ”Numerical Simulation of Stochastic Diffe-
tervals for individual samples (solid green lines), marking a rential Algebraic Equations for Power System Transient Stability with
measure of stochastic trajectories dispersion. In case of the Random Loads,” in Proc. 2011 IEEE Power and Energy Society General
Meeting, San Diego, pp. 1-8, 2011.
additive noise the resultant stochastic processes keep their [7] A. Zjajo, Q. Tang, M. Berkelaar, J. P. de Gyvez, A. Di Bucchianico,and
Gaussian distribution, and a student-t distribution can be N. van der Meijs, ”Stochastic Analysis of Deep-Submicrometer CMOS
directly applied. In case of the multiplicative noise, however, Process for Reliable Circuits Designs,” IEEE Transactions on Circuits
and Systems I: Regular Papers, vol. 58, no. 1, pp. 164-175, 2011.
this is not valid, and the sample statistics are calculated via [8] E. Kolářová, ”Modelling RL Electrical Circuits by Stochastic Diferential
dividing the samples set into a number of batches while Equations,” in Proceedings of International Conference on Computer as
utilizing a central limit theorem, see [3] for more details. The a Tool, Belgrade (Serbia and Montenegro): IEEE R8, pp. 1236-1238,
2005.
noise intensities were chosen as α = 10−6 in case of the noisy [9] E. Kolářová, ”Statistical estimates of stochastic solutions of RL electrical
source vS∗ , and αk = 10−5 , ∀k, in case of the noisy series circuits,” in Proc. of the IEEE International Conference on Industrial
resistances Rd∗ . Distinct noise intensities for different noisy Technology ICIT06, Mumbai, pp. 2546-2550, 2006.
[10] N. S. Patil, B. G. Gawalwad, and S. N. Sharma, ”A random input-
elements, which result in comparable dispersions of stochastic driven resistor-capacitor series circuit,” in Proc. 2011 International
processes, can help to evaluate components of the TL model Conference on Recent Advancements in Electrical, Electronics and
with significant noise effects. It should be noticed that stochas- Control Engineering, Sivakasi, pp. 100-103, 2011.
[11] F. Rahman, and N. Parisa, ”A stochastic perspective of RL electrical
tic processes dWk (t), needed for all numerical solutions, are circuit using different noise terms,” COMPEL: The International Journal
generated by the MATLABr function for normally distributed for Computation and Mathematics in Electrical and Electronic Engi-
random numbers. neering, vol. 30, no. 2, pp. 812-822, 2011.
[12] C. K. Cheng, J. Lillis, S. Lin, and N. Chang, Interconnect Analysis and
Synthesis, New York: John Wiley & Sons, 2000.
VI. C ONCLUSION [13] M. P. Li, Jitter, Noise, and Signal Integrity at High-Speed, Upper Saddle
Real implementations of the interconnects in high-speed River: Prentice Hall, 2007.
[14] E. Kolářová, and L. Brančı́k, ”Application of stochastic differential
electronic systems are being influenced by a number of equations in second-order electrical circuits analysis,” Przeglad Elek-
physical effects [13]. The stochastic differential equations trotechniczny, vol. 88, no. 7a, 2012, pp. 103-107.
approach can therefore be an interesting alternative to other [15] L. Brančı́k, ”Time and Laplace-domain methods for MTL transient and
sensitivity analysis,” COMPEL: The International Journal for Compu-
probabilistic approaches, how to process such the random tation and Mathematics in Electrical and Electronic Engineering, vol.
effects. We can mention at least [18], where a stochastic 30, no. 4, pp. 1205-1223, 2011.
Galerkin method for the solution of stochastic telegrapher’s [16] L. Brančı́k, and E. Kolářová, ”Analysis of higher-order electrical circuits
with noisy sources via stochastic differential equations approach,” in
equations is used, [19], where a polynomial-chaos expansion Proc. of 22nd Int. Conf. Radioelektronika’2012, Brno, 2012, pp. 81-84.
method is adopted, or [20], where a method based on a [17] E. Kolářová, and L. Brančı́k, ”Vector linear stochastic differential
classical Monte Carlo analysis is considered. In mentioned equations and their applications to electrical networks,” in Proc. of
35th Int. Conf. on Telecommunications and Signal Processing TSP2012,
cases, however, the solutions are performed in the frequency Prague, 2012, pp. 311-315.
domain, in contrast to our work, where the responses in the [18] D. V. Ginste, D. De Zutter, D. Deschrijver, T. Dhaene, P. Manfredi,
time domain are directly treated. Especially at the solution and F. Canavero, ”Stochastic Modeling-Based Variability Analysis of
On-Chip Interconnects”, IEEE Trans. on Components, Packaging and
of signal integrity issues an overall dispersion in the time Manuf. Technology, vol. 2, no. 7, 2012, pp. 1182-1192.
domain can help to evaluate whether digital signals delivered [19] D. Spina, F. Ferranti, T. Dhaene, L. Knockaert, G. Antonini, and D. V.
are treatable by receivers. However, the application of the SDE Ginste, ”Variability analysis of multiport systems via polynomial-chaos
expansion”, IEEE Trans. on MTT, vol. 60, no. 8, 2012, pp. 2329-2338.
theory in the TLs modelling in the frequency domain is also [20] D. V. Ginste, D. D. Zutter, D. Deschrijver, T. Dhaene, and F. Canavero,
enabled. ”Macromodeling based variability analysis of an inverted embedded
We will focus our further research on a few directions. microstrip line” in Proc. IEEE 20th Int. Conf. Elect. Perform. Electron.
From the point of view of theoretical studies the remaining Packag. Syst., San Jose, CA, 2011, pp. 153-156.
[21] L. Brančı́k, and E. Kolářová, ”Stochastic differential equations approach
RLCG circuit elements will be taken into account to produce in the analysis of MTLs with randomly varied parameters,” in Proc.
stochastic effects, and also their simultaneous actions will be of 19th IEEE International Conference on Electronics, Circuits, and
investigated. A SDEs approach will again be used to model Systems ICECS2012, Sevilla, 2012, pp. 725-728.
higher-order systems, especially those with multiconductor
transmission lines (MTL) [15], [21]. Finally, other higher-
order numerical techniques consistent with the Itô stochastic
calculus will further be considered to ensure even better
stability and accuracy of the solutions.