15.093J/2.

098J Optimization Methods

Professor: Dimitris Bertsimas

1 Structure of Class
Network Flows: Lec. 10-11

Linear Optimization (LO): Lec. 1-9

Slide 1

Discrete Optimization: Lec. 12-15 Dynamic Optimization: Lec. 16 Nonlinear Optimization (NLO): Lec. 17-24

2 Requirements
Homeworks: 30% Midterm Exam: 30% Final Exam: 40% Class Participation: important tie breaker Use of commercial software for solving optimization problems

Slide 2

3 Lecture Outline
History of Optimization Where LOPs Arise? Examples of Formulations

Slide 3

4 History of Optimization
Fermat, 1638 Newton, 1670
min f (x) x: scalar df (x) = 0 dx 1

Slide 4

Euler, 1755 Lagrange, 1797

min f (x1 : : : xn) rf (x) = 0 min f (x1 : : : xn) s.t. gk (x1 : : : xn) = 0 k = 1 : : : m

Euler, Lagrange Problems in in nite dimensions, calculus of variations.

5 Nonlinear Optimization
5.1 The general problem

min f (x1 : : : xn) s:t: g1(x1 : : : xn) 0 ... gm (x1 : : : xn) 0:

Slide 5

6 What is Linear Optimization?

6.1 Formulation
minimize 3x1 + x2 subject to x1 + 2x2 2 2x1 + x2 3 x1 0 x2 0 2 1 x= x b= 3 x2 minimize c x subject to Ax b

Slide 6

c= 3 1

2 A= 1 2 1

x 0

7 History of LO
lute values.

7.1 The pre-algorithmic period

Fourier, 1826 Method for solving system of linear inequalities. de la Vallee Poussin simplex-like method for objective function with abso2

Slide 7

Kantorovich, Koopmans, 1930s Formulations and solution method. von Neumann, 1928 game theory, duality. Farkas, Minkowski, Caratheodory, 1870-1930 Foundations.

7.2 The modern period

George Dantzig, 1947 Simplex method. 1950s Applications. 1960s Large Scale Optimization. 1970s Complexity theory. Khachyan, 1979 The ellipsoid algorithm. Karmakar, 1984 Interior point algorithms.

Slide 8

8 Where do LOPs Arise?

8.1 Wide Applicability
Transportation
Slide 9

Air tra c control, Crew scheduling, . . . Movement of Truck Loads

Telecommunications Manufacturing Medicine Engineering ATEX) Typesetting (TEX, L

9 Transportation Problem
9.1 Data
m plants, n warehouses si supply of ith plant, i = 1 : : :m dj demand of j th warehouse, j = 1 : : :n cij : cost of transportation i ! j

Slide 10

9.2 Decision Variables

9.2.1 Formulation
xij = number of units to send i ! j m n XX min cij xij i=1 j =1 m X s:t: xij = dj i=1 n X xij = si
j =1 xij
Slide 11

j = 1 : : :n i = 1 : : :m

0
Slide 12

10 Sorting through LO

Given n numbers c1 c2 : : : cn The order statistic c(1) c(2) : : : c(n): c(1) c(2) : : : c(n) P Use LO to nd k i=1 c(i) . n X min ci xi i=1 n X s:t: xi = k i=1 0 xi 1 i = 1 : : : n

11 Investment under taxation

You have purchased si shares of stock i at price qi, i = 1 : : : n. Current price of stock i is pi You expect that the price of stock i one year from now will be ri. You pay a capital-gains tax at the rate of 30% on any capital gains at the time of the sale. You want to raise C amount of cash after taxes. You pay 1% in transaction costs. Example: You sell 1,000 shares at $50 per share you have bought them at $30 per share Net cash is: 50 1 000 ; 0:30 (50 ; 30) 1 000; 0:01 50 1 000 = $43 500:

Slide 13

11.1 Formulation
max s:t:
n X i=1 n

ri (si ; xi )

Slide 14

n n X X pi xi ; 0:30 (pi ; qi)xi ; 0:01 pi xi C i=1 i=1 i=1 0 xi si

12 Investment Problem
Five investment choices A, B, C, D, E. A, C, and D are available in 1993. B is available in 1994. E is available in 1995. Cash earns 6% per year. $1,000,000 in 1993. Year: 1993 1994 1995 1996 LIMIT

Slide 15

12.1 Cash Flowper Dollar Invested

A B C -1.00 +0.30 +1.00 0 $500,000 0 -1.00 +0.30 +1.00 NONE -1.00 +1.10 0 0 $500,000

D -1.00 0 0 +1.75 NONE

E 0 0 -1.00 +1.40 $750,000

Slide 16

12.2 Formulation

12.2.1 Decision Variables

A : : :E : amount invested in $ millions Casht : amount invested in cash in period t, t = 1 2 3 max 1:06Cash3 + 1:00B + 1:75D + 1:40E s:t: A + C + D + Cash1 1 Cash2 + B 0:3A + 1:1C + 1:06Cash1 Cash3 + 1:0E 1:0A + 0:3B + 1:06Cash2 A 0:5 C 0:5 E 0:75 A ::: E 0 5

Slide 17

Solution: A = 0:5M , B = 0, C = 0, D = 0:5M , E = 0:659M , Cash1 = 0, Cash2 = :15M , Cash3 = 0 Objective: 1:7976M

13 Manufacturing
13.1 Data

n products, m raw materials cj : pro t of product j bi : available units of material i. aij : # units of material i product j needs in order to be produced.

Slide 18

13.2 Formulation

13.2.1 Decision variables

xj = amount of product j produced. n X max cj xj

j =1

Slide 19

s:t: a11x1 + + a1nxn b1 .. . am1 x1 + + amn xn bm xj 0 j = 1 : : :n

14 Capacity Expansion
14.1 Data and Constraints
Dt : forecasted demand for electricity at year t Et : existing capacity (in oil) available at t ct : cost to produce 1MW using coal capacity nt : cost to produce 1MW using nuclear capacity No more than 20% nuclear Coal plants last 20 years Nuclear plants last 15 years

Slide 20

14.2 Decision Variables

xt : amount of coal capacity brought on line in year t. yt : amount of nuclear capacity brought on line in year t. wt : total coal capacity in year t. zt : total nuclear capacity in year t.

Slide 21

14.3 Formulation
min s:t:

ct xt + ntyt t=1 t P wt = xs s=max(0 t;19) t P zt = ys s=max(0 t;14) wt + zt + Et Dt zt 0:2(wt + zt + Et) xt yt wt zt 0:

T P

Slide 22

t = 1 : : :T t = 1 : : :T

15 Scheduling

15.1 Decision variables

Decision Variables

Hospital wants to make weekly nightshift for its nurses Dj demand for nurses, j = 1 : : : 7 Every nurse works 5 days in a row Goal: hire minimum number of nurses

Slide 23

xj : # nurses starting their week on day j

15.2 Formulation
7 P min xj j =1 s:t: x1+ x1+ x1+ x1+ x1+ xj 0

Slide 24

x2 x2 + x2 + x2 + x2 +

x3 x3 + x3 + x3 + x3 +

x4 + x5 + x6 + x5 + x6 + x6 + x4 + x4 + x5 x4 + x5 + x6 x4 + x5 + x6 + 7

x7 x7 x7 x7 x7

d1 d2 d3 d4 d5 d6 d7

16 Revenue Management
16.1 The industry

Deregulation in 1978 Prior to Deregulation { Carriers only allowed to y certain routes. Hence airlines such as Northwest, Eastern, Southwest, etc. { Fares determined by Civil Aeronautics Board (CAB) based on mileage and other costs (CAB no longer exists)

Slide 25

Post Deregulation anyone can y, anywhere fares determined by carrier (and the market)

Slide 26

17 Revenue Management
17.1 Economics

Huge sunk and xed costs Very low variable costs per passenger ($10/passenger or less) Strong economically competitive environment Near-perfect information and negligible cost of information Highly perishable inventory Result: Multiple fares

Slide 27

18 Revenue Management
18.1 Data

n origins, n destinations 1 hub 2 classes (for simplicity), Q-class, Y-class Q rY Revenues rij ij Capacities: Ci0, i = 1 : : :n C0j , j = 1 : : :n Q , DY Expected demands: Dij ij

Slide 28

18.2 LO Formulation

Qij Yij

18.2.1 Decision Variables

maximize subject to

: # of Q-class customers we accept from i to j : # of Y-class customers we accept from i to j

Slide 29

X X
i j n

rij Qij

+r
ij

Y ij Yij

X
j n i

(Q + Y ) C 0
ij i

0Q

(Q + Y ) C0
ij ij ij Q Dij

j Y Dij

0Y

ij

19 Revenue Management
19.1 Importance
Robert Crandall, former CEO of American Airlines: We estimate that RM has generated $1.4 billion in incremental revenue for American Airlines in the last three years alone. This is not a one-time bene t. We expect RM to generate at least $500 million annually for the foreseeable future. As we continue to invest in the enhancement of DINAMO we expect to capture an even larger revenue premium.
Slide 30

20 Messages

20.1 How to formulate?

1. De ne your decision variables clearly. 2. Write constraints and objective function. 3. No systematic method available.

Slide 31

What is a good LO formulation?

A formulation with a small number of variables and constraints, and the matrix A is sparse.

21 Nonlinear Optimization
21.1 The general problem

min f (x1 : : : xn) s:t: g1(x1 : : : xn) 0 ... gm (x1 : : : xn) 0:

Slide 32

22 Convex functions
f : S ;! R For all x1 x2 2 S f ( x1 + (1 ; )x2 ) f (x) concave if ;f (x) convex. f (x1 ) + (1 ; )f (x2 )

Slide 33

23 On the power of LO
23.1 LO formulation

min f (x) = maxk dk x + ck s:t: Ax b min z s:t: Ax b dk x + ck z 8 k

0

Slide 34

24 On the power of LO
24.1 Problems with
Idea: jxj = maxfx ;xg
j:j

min s:t:

P c jx j

Ax b
j j

j j

Slide 35

xj zj ;xj zj Message: Minimizing Piecewise linear convex function can be modelled by LO 10

min s:t:

Pc z

Ax b

MIT OpenCourseWare http://ocw.mit.edu

15.093J / 6.255J Optimization Methods

Fall 2009

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.