Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
= (D(p(t)) S(p(t)))
with D(p) and S(p) being the demand and supply at price p respectively, and > 0 some constant. This equation obviously assumes
that the supply and demand depend only on the price at the current
time t. This may be reasonable in a static model, where we only have
the present time. However, agents in a dynamic market may develop
expectation regarding the price, according to its past trend. The idea
is that agents may assume price convey information on other agents
behaviour (and they may be right doing so). So, for example, we may
have agents who look at the price, and see that the price is going up
(p > 0), and going up quicker every minute (
p > 0) and so may change
their demand accordingly. So it may be that in this case change in
price is again due to change in excess demand, only now, demand and
supply are functions of p, p,
and p (the second derivative w.r.t time of
p), hence we will get the equation
(1.1.1)
p(t)
= (D(p(t), p(t),
p(t)))
t).
= B. A solution
will be determined by the initial conditions x(0) = x0 and x(0)
= x 0
(or, more generally, x(t0 ) and x(t
0 ) for some initial time t0 ).
1.2. Reducible equations. These are equations which are independent of x, namely x(t) = F (x(t),
we obtain
1
the equation
(1.2.1)
z(t)
= F (z(t), t)
which is a 1st order equation. If we are able to solve this equation (by
the methods we introduced above), then we can write
Z
(1.2.2)
x(t) = z(t)dt.
Example 1.2.1. x = x.
Denote z = x.
Then we obtain the equation
z = z z(t) = Aet ,
(1.2.3)
R
x + a(t)x + b(t)x = 0.
The term linear follows from the fact that if x1 (t) and x2 (t) satisfy
equation (1.3.1) then so does rx1 (t) + sx2 (t) for every r, s R. Indeed,
if r, s R and x1 (t) and x2 (t) satisfy equation (1.3.1) then
d2
d
(rx1 + sx2 ) + a(t) (rx1 + sx2 ) + b(t)(rx1 + sx2 ) =
2
dt
dt
r(
x1 + a(t)x 1 + b(t)x1 ) + s(
x2 + a(t)x 2 + b(t)x2 ) = r 0 + s 0 = 0.
Theorem 1.3.1. If x1 and x2 are linearly independent solutions of
equation (1.3.1) then every solution x of equation (1.3.1) has the form
x = rx1 + sx2 for some r, s R, with r, s R determined by the initial
conditions.
Theorem 1.3.1 states, in fact, that the general solution of equation
(1.3.1) is of the form x(t) = Ax1 (t) + Bx2 (t), where x1 , x2 are some
linearly independent solutions of equation (1.3.1).
Example 1.3.2. Consider the equation x + x = 0. What is the general
solution of this equation? Theorem 1.3.1 states that it is sufficient to
find two linearly independent solutions. Can we think of functions that
satisfy x + x = 0? We shall use a trick - multiply the equation by 2x.
We thus obtain 2
xx + 2xx
= 0. Integration w.r.t. t yields:
Z
(1.3.2)
C = (2
xx + 2xx)dt
= x 2 + x2 .
MATH ECON II
1.4. Homogeneous equations with constant coefficients. Suppose that a(t) = a and b(t) = b are constants. We wish to find the
general solution of the equation
(1.4.1)
x + ax + bx = 0.
(1.4.2)
a2
4
Theorem 1.4.1.
1. If > 0 then r1 6= r2 are real numbers, and
the general solution x to equation (1.4.1) is given by:
(1.4.4)
x(t) = (A + Bt)ert .
at
x(t) = (A + Bt)ert ,
x(t) = (A + Bt)et .
(1.4.12)
p = D(p, p,
p) S(p, p,
p) =
D0 (t) S0 (t) ( + a)p + ( + b)p + ( + c)
p.
MATH ECON II
x + ax + bx = f (t),
= 0, and 2. finding a
particular solution for the equation (1.4.14). As we already know how
to find the general solution for the homogeneous equation, we shall now
explore methods for finding a particular solution in various cases.
Example 1.4.7. f (t) = C, with C being a constant. We will be looking
for a particular solution xp (t) = B with B being a constant. It is easy
to verify that the particular solution is in fact xp (t) = Cb .
n
P
Example 1.4.8. f is a polynomial. Namely, f (t) =
ci ti , with c0 , c1 , ..., cn
i=0
d2 n
R. Notice that dtd (tn ) = ntn1 and dt2
(t ) = n(n1)tn2 . This leads us
to look for a polynomial particular solution. Namely, we assume that
n
P
xp (t) =
di ti , where the degree n of the polynomial xp is the same as
i=0
= 0 is given
by (verify that!)
xh (t) = (A + Bt)et ,
and thus, the general solution for the nonhomogeneous equation x +
2x + x = t2 is (by Theorem 1.4.6)
x(t) = (A + Bt)et + t2 4t 10.
Example 1.4.9. f (t) = A exp(t). We shall look for particular solution
of the form
(1.4.15)
xp (t) = B exp(t).
Example 1.4.10. f (t) = A sin(t) + B cos(t). We shall look for a solution xp (t) = C cos(t) + D sin(t). For example, consider the equation
x + x + x = cos t. Here = 1, and we are looking for a particular solution of the form xp (t) = C cos t + D sin t. Thus x p (t) = D cos t C sin t,
and xp (t) = C cos t D cos t = xp (t). Therefore, by substituting
back into the equation we obtain:
cos t = xp + x p + xp = xp (t) + x p + xp = x p (t) = D cos t C sin t
(1.4.17)
Now, the only way to obtain an equality above, the coefficient of sin on
both sides of the equation should be equal ,and the same should hold
for cos. Therefore, D = 1 and C = 0 and we find that xp (t) = cos t is
a particular solution of the equation.
1.5. Market Model with Price Expectations. Lets go back to the
example of price adjustment with speculation and fluctuation. Suppose
D0 (t) = F0 cos(t), S0 (t) = 0, and that = 1 + a = k b + b = d + 1,
+ c = m Then the price adjustment equation transforms into
(1.5.1)
F0 cos(t) = kp + dp + m
p
MATH ECON II
For the particular solution we look for a function of the form pp (t) =
A cos(t) + B sin(t). Then pp (t) = A sin(t) + B cos(t) and
pp (t) = A 2 cos(t) B 2 sin(t) = 2 pp (t). So
F0 cos(t) = kp + dp + m
p = kp + dp m 2 p = dp + (k m 2 )p =
d(B cos(t) A sin(t)) + (k m 2 )(A cos(t) + B sin(t))
(k m 2 )A + dB = F0 , dwA + (k m 2 )B = 0
A=
(k m 2 )F0
dF0
, B=
2
2
2
(k m ) + (d)
(k m 2 )2 + (d)2
2
d
Let 1 = [(kmkm
Notice that
2 )2 +(d)2 ]1/2 and 2 = [(km 2 )2 +(d)2 ]1/2 .
2
2
1 +2 = 1, so (1 , 2 , 1) is a Pythagorean triplet (i.e., these three nonnegative numbers obey
the Pythagorean theorem). We thus deduce
that there is a 0, 2 s.t. 1 = cos and 2 = sin . We can now
write the particular solution of the equation as
F0
(cos cos(t) + sin sin(t)) =
pp (t) =
2
[(k m )2 + (d)2 ]1/2
F0
(1.5.2)
cos(t ),
2
[(k m )2 + (d)2 ]1/2
where the last equality above follows from the trigonometric identity
cos(x + y) = cos x cos y sin x sin y. So prices oscillate with amplitude
F0
(after the introduction of a phase). What happens as
(km 2 )2 +(d)2
2
0
k m approaches 0? In this case the amplitude [(km2 )F2 +(d)
2 ]1/2
F0
k
2
will increase and in the limit m it will approach d .
w = f (U )
with f being some monotonically decreasing function (so more unemployment means slower growth rate w). A similar relation, baring the
same name, can be written for the empirical relationship between unemployment and the rate of inflation p, which is the rate of growth
of price index P . The relating economic factor between the two (w
and p) is the labor productivity `, which is given exogenously, and for
simplicity assume p = w ` (so the higher productivity is the lower
the inflation is, as you may produce the demand for lower costs, and if
wages increase sharply it will create inflation). So we can now write
(1.6.2)
p = f (U ) `.
(t)
= (p(t) (t)).
(t)
=
(b(p(t) m(t)) + (t)
` (t)),
which is a differential equation of the form
(1.6.9)
+ A + B = f (t)
Given `, a
with A = b , B = b, and f (t) = bm(t) `.
central bank can control m(t) and thus influence inflation expectations,
1Notice, that a first possible tweak of the model would be to account for indirect changes, like
changing wages for hour, shorter hours, seasonal workers etc.
MATH ECON II
and through that (given that we solve the second order equation (1.6.9)
affect inflation rates (Equation 1.6.4) and unemployment (1.6.5).