Dynamic Optimization

Please review the following exercises.

You may only solve those you feel not comfortable with.

Integration Exercises
1. Evaluate the following:
a)
b)
c)
d)
e)
f)
g)

16x dx
9x dx
x 3x dx
2e dx
4xe dx
5e 3 / x dx
2x
x 3 dx
3

h)
i)

x2 2

j)

dx

x dx
2

1
3

k)

2x

2
4

1/ 2

2x

x 3 x 1
x ln(x)dx

l)

x
2

e x dx
6x 2 dx

2. Given the following marginal revenue functions:

a)
b)

MR(q) = 28q-e0.3q
MR(q) = 10(1+q)-2

find in each case the total revenue function.

3. Given the marginal propensity to import M(Y) = 0.1, find the import function M(Y)
when M = 20 if Y = 0.
4. Find the consumption function C(Y), given the marginal propensity to consume
C '(Y) 0.8 0.1/ Y where C(100) = 100.
5. Given the marginal cost MC(q) 3q 2 18q 30 , determine the decrease in total cost C(q)
as output is reduced from 12 units to 3 units.
6. If the marginal propensity to save is given by S'(Y) 1 1.5 / Y and aggregate saving is
zero when income is 100, find the saving function.
7. If the marginal propensity to consume is given by C(Y) = 4800(2+Y)-3 and aggregate
consumption is 400 when income is zero, find the consumption function, C(Y).
8. Given the marginal cost MC(q) = 7.5e0.15q and fixed cost is 80, find the total cost function.
9. Assume that the rate of investment is given by I(t) = 12t1/3, where K(0) = 25: a) Find the
time path of capital stock K. b) Find the amount of capital accumulation during the time
intervals [0,1] and [1,3]. Hint: dK/dt=I(t).

Steady State Analysis

1) Solve the following systems, sketch a phase diagram and discuss the qualitative behavior of
solutions to each of these systems:
a)
dx/dt = 3x-2y
dy/dt = 2x-2y
Solution sketch
3 2
A

2 2
Characteristic equation

0 2 2
Solve for eigenvalues
1 1

2 2
Solve for eigenvector
Ak k

3 2 k1
2 2 k

For 1 1

k1

k2

-2

-1

k12 2k11

-1

Let k11 1,
k11
1
then 1
2
k 2
For 2 2
k12 2k 22
Let k 22 1,
k12
2

2
1
k 2
GeneralSolution
then

1
2
x(t)
1t k1
2 t k1

c
e

c
e
1
2
1
2
y(t)

k 2
k2
x(t) c1e t 2c 2e 2t

y(t) 2c1e t c 2 e 2t
The critical point at the origin is an unstable saddle point

-2

b)
dx/dt = -2x+y
dy/dt = x-2y

Solution
The critical point at the origin is a stable
improper node
x(t) c1e t c 2e 3t

y(t) c1e t c2 e 3t

-2

-1

-1

-2

c)
dx/dt = x-y
dy/dt = 5x-3y
Solution
The critical point at the origin is a stable spiral
point
x t e t c1 cos t c 2 (2 cos t sin t)
y t e

-2

-1

c1 sin t c 2 (cos t 2sin t)

-1

-2

2) Solve these initial value problems

a)
dx/dt = 9x+5y
dy/dt = -6x-2y
x(0) = 1, y(0) = 0
Solution sketch
9 5
A

6 2
0 2 7 12
1 4 k12 k11
6
2 3 k 22 k12
5
Generalsolution
1
2
x(t)
1t k1
2 t k1

c1e 1 c 2 e 2
y(t)
k 2
k2
x(t)
4t 1
3t 5

c1e c 2 e
y(t)
1
6

x(t) c1e 4t 5c 2e3t

y(t) c1e 4t 6c 2e3t
Particular solution
c1 6
c2 1
x(t) 6e 4t 5e3t
y(t) 6e 4t 6e3t

b)
dx/dt = x-5y
dy/dt = x-3y
x(0) = y(0) = 1
Solution sketch
1 5
A

1 3
0 2 2 2
1 1 i (2 i)k11 5k12
Let k11 5 k12 2 i
2 1 i (2 i)k12 5k 22
Let k12 5 k 22 2 i
Generalsolution
x(t) 5e( 1i)t
x(t) 5e t cos t i 5e t sin t
x(t) c1 realpart x c 2 imaginarypart x
x(t) c1 5e t cos t c 2 5e t sin t
y(t) (2 i)e ( 1i)t i e t (cos t i sin t)
y(t) e t (2 cos t sin t) i e t (2sin t cos t)
y(t) c1 realpart y c 2 imaginarypart x
y(t) c1e t (2 cos t sin t) c 2e t (2sin t cos t)
Particular solution
c1 0.2
c 2 0.6
x(t) e t (cos t 3sin t)
y(t) e t (cos t sin t)

3) For each of the following non-linear systems of excess demand equations, determine
equilibrium prices, assess the stability properties of those prices and sketch a phase diagram
in px py space. You may assume a Walrasian price adjustment; i.e. (i) dpx/dt = Ex(px,py) and
(ii) dpy/dt = Ex(px,py).
a)
Ex(px,py) = 50 6px + 2py
Ey(px,py) = 35 3px 4py
Solution sketch
p*x , p*y 9, 2
Let

p%x p x p*x
p%y p y p*y

Then

dp x dt dp%x dt
dp y dt dp%y dt

Thus

dp%x dt 6p%x 2p%y

dp%y dt 3p%x 4p%y

Solve this
6 2

3 4

tr(A) 10
2

tr(A)
det(A) 30 25

2
The characteristic equation has complex
but not pure imaginary roots.
We have a stable spiral point.
Draw phase diagram.

10

11

b)
Ex(px,py) = 80 4px + 2py
Ey(px,py) = 50 + 2px - 6py
20

19

18

17

28

Solution
Stable Improper Node.

29

30

31

c)
Ex(px,py) = -5 2px + 3py
Ey(px,py) = 16 + 4px - 8py

4.5
4
3.5
3
2.5
2
1.5

Solution
Stable Improper Node.

d)
Ex(px,py) = -13 + 2px + 3py
Ey(px,py) = 16 + 4px -8py

4.5
4
3.5
3
2.5
2
1.5

Solution
Unstable saddle point.

4) Interpret each of the following non-linear systems as describing the interaction of two
competing species with populations x and y. determine the critical points (equilibrium points),
identify the type and stability properties of each, and sketch a phase diagram to illustrate
your results.
a)
dx/dt = x(1.5-x-0.5y)
dy/dt = y(2-y-0.75x)
Solution sketch
Equilibrium points

x, y 0, 0 , 0, 2 , 3 2, 0 , 4 5, 7 5
Linearize
dx
3
1
1
x, y : 2x * y* x x* x* y y*
dt
2
2
2

dy
3
3
x, y : x* x x* 2 2y* x* y y*
dt
4
4

Evaluate linearized system for all critical points

1
1
3

2x * y*
x*

2
2
A x * , y* 2

3 *
3 *
*

x
2 2y x

4
4
3
0
A 0, 0 2 unstable spiral point

0 2
1

0
2
A 0, 2
unstable saddle point
3 2

3
3

3
2
4
A , 0
unstable saddle point
7
2
0

2
4

4
7

5
5
A ,
stable improper node
5 5
3 7

5
5

b)
dx/dt = x(1.5-x-0.5y)
dy/dt = y(2-0.5y-1.5x)
Equilibrium points

x, y 0, 0 , 0, 4 , 3 2, 0 , 1,1
Linearize
dx
3
1
1
x, y : 2x* y* x x* x* y y*
dt
2
2
2

dy
3
3
x, y : x* x x* 2 y* x* y y*
dt
2
2

Evaluate linearized system for all critical points

1
1
3

2x* y*
x*

2
2
A x * , y* 2

3 *
3 *
*

x
2 y x

2
2
3
0
A 0, 0 2 unstable spiral point

0 2

0
stable improper node

2
3
3

3
4 stable improper node
A , 0 2

2
0 1

4
1

1 2
A 1,1
unstablesaddle point
2 1

2
2
1

A 0, 4 2

Calculus of Variation Exercises

Solve the following problems:
1
T

Max x '2 dt
0

x 0 0

s.t.

x T B
2
t1

Max x '2 10tx dt

t0

x t0 x0

s.t.

x t1 x1
3
t1

Max tx ' x '2 dt

t0

x t0 x0

s.t.

x t1 x1
4
Optimal Production Inventory path
Production cost rise linearly with level of production
Storage cost are constant per unit or production
T

Max c1x '2 c 2 x dt

0

s.t.

x t0 0

x t1 B
5
Consumption Capital investment path
Maximize utility over time
Utility comes from consumption
Capital returns plus wages can be used for consumption and increase in capital holdings
T

Max e rt U C t dt
0

s.t.

U C t ln C

iK t w t K ' t C t
K t0 K0
K T 0

Optimal Control Exercises

1. The island economy of Utopia is based on the harvest of a single resource by native divers.
The island chief is concerned about the resource stock left to unborn generations and has
solicited your advice in resource management. After discussions with the chief, as well as
some biological research, you determine the following:
U(y(t)) = y(t ) 0.75 = utility of yield y(t)
2

x (t) = x(t) - 0.001x(t ) - y(t) = equation of motion

= 0.1 = the tribal discount rate

3.

a)

State the current value Hamiltonian?

b)

Solve for the steady state optimum (x, y, m). Note, steady state is where all changes
with respect to time equal zero.

A mine owner possesses resource stock, S, extractable at a constant cost per unit of c. The
owner wishes to maximize discounted profits over a finite time horizon [0,T]. The rate of
production is y(t) and the price is given by the inverse demand function p(t) = g(y(t)),
which is assumed to satisfy the normal conditions placed on inverse demand function. Find
the optimal path for the depletion of the resource stock. Interpret the necessary conditions
for this problem along with the path assuming g(y(t)) = a - by(t):
T

max (y(t) g(y(t) ) - c y(t) ) e -rt dt

0

s.t .
y(0) = 0
y(T) = S.
Hint: A function whose derivative equals zero must itself be a constant, therefore there is no
need to carry out the differentiation. Write out the necessary conditions before taking all the
derivatives.
4.

One of the first examples of OC is the derivation of capital accumulation for a one sector
economy with an exponentially growing population and production under constant returns to
scale. The social objective for a society with time horizon T is to maximize:
T

W = e (n-r) t u(c) dt
0

where r is the social rate of time preference, per capita consumption is denoted by c, the
utility enjoyed by a person consuming at rate c is given by u(c), n denotes the population
growth rate, and r> n. Let the change in capital per capita be k(t) = f(k) - c - (n+)k, where
f(k) is a function independent of c, and is the rate of decay of the physical stock. The
productivity of capital is given by f(k). Interpret the optimality and costate conditions for
this problem. Give the function for and interpret the meaning of the costate multiplier.

The Neoclassical Growth Model

Assumptions
Constant population of L identical workers

per capita consumption = c(t)

Each worker derives utility U(c) from consumption with U>0 and U<0

Lim(c0)U(c) =

rt
Central planner tries to maximize e LU(c(t))dt subject to production possibilities of the
0

economy.
Production function for a single good in the economy is F(K,L) and homogeneous of degree 1.
F(0,L) = F(K,0) = 0, FK > 0, FKK < 0, FL > 0, FLL < 0
Produced goods can be consumed or invested in new capital F(K,L) = Lc + K
K(0) = K0

Analysis
The dynamic optimization problem is set up as follows

max K L e rt U(F(K, L) K ')dt

0

s.t.

K(0) K 0 , K 0, F(K, L) K ' 0

To simplify the analysis, let k = K/L and f(k) = F(K/L,1). Under the assumption of homogeneity of
degree 1, we have F(K,L) = L*F(K/L,L/L) = L*f(K) where f(0) = 0, f(k) > 0, and f(k) < 0. (Note
that 0 FK Lf '(k)k / K Lf '(k) / L f '(k) and 0 FKK f ''(k)k / K f ''(k) / L ). Algebraic
manipulation yields k=f(k)-c. This leads to

max k L e rt U(f (k) k ')dt

s.t.

c = 0

k(0) k 0 , k 0, f (k) k ' 0

The Euler equation for this problem is

dU '(c) / dt / U '(c) U ''(c)c '/ U ' f '(k) r

k = 0

Phase diagram:
c = 0 if f(k) = r
k = 0 if c = f(k) with f(k) > 0, and f(k) < 0 from above
The resulting phase diagram is shown to the right.
From the graph it is easy to see that the steady state is
an unstable saddle point. This can be verified through an
k
algebraic analysis where the behavior of the system is
approximated by a related system of linear differential
equations (use Taylor expansion and disregard h.o.t):
U '(css )
f ''(k ss )(k k ss ) with the characteristic equation:
k ' f '(k ss )(k k ss ) (c css ) and c '
U ''(css )
0 2 f '(k ss ) f ''(k ss )

U '(css )
with 1,2 f '(k ss )
U ''(css )

f '(k ss )

It is easy to see that the roots are real and of opposite sign.

4f ''(k ss )

U '(css )
U ''(css )

5.

Questions to the Neoclassical Growth Model

From Kamien and Schwartz (Second edition, p.183)
Show that if, in the Neoclassical Growth Model utility U(c,k) depends on capital stock
(wealth), as well as on consumption, Uc > 0, Ucc < 0, Uck = 0, Uk > 0, Ukk < 0, then there
may be multiple equilibriums. Identify the nature of these saddle points.

6.

A population of N fish in a certain lake grow at rate N(t) = aN(t) bN2(t) if undisturbed by
people. Fish can be withdrawn by the lake and consumed at rate c(t), yielding utility u(c(t))
to the consuming community and reducing the fish growth rate accordingly: N(t) = aN(t)
bN2(t) c(t). Assume future utilities to the community are discounted at constant rate r.
Characterize a fishing (consumption) plan to maximize the present value of the discounted
stream of utilities. Assume that N(0) = a/b, u > 0, and u < 0.