DESIGN AND PLANNING

UNDER UNCERTAINTY
Prof. Miguel Bagajewicz
CHE 4273

Two
-Stage
Two-Stage
Stochastic Optimization Models
Philosophy
Maximize the Expected Value of the objective over all possible realizations of
uncertain parameters.
Typically, the objective is Profit or Net Present Value.
Sometimes the minimization of Cost is considered as objective.

Uncertainty
Typically, the uncertain parameters are: market demands, availabilities,
prices, process yields, rate of interest, inflation, etc.
In Two-Stage Programming, uncertainty is modeled through a finite number
of independent Scenarios.
Scenarios are typically formed by random samples taken from the probability
distributions of the uncertain parameters.
2

Characteristics of Two
-Stage
Two-Stage
Stochastic Optimization Models

First-Stage Decisions
Taken before the uncertainty is revealed. They usually correspond to structural
decisions (not operational).
Also called Here and Now decisions.
Represented by Design Variables.
Examples:
To build a plant or not. How much capacity should be added, etc.
To place an order now.
To sign contracts or buy options.
To pick a reactor volume, to pick a certain number of trays and size
the condenser and the reboiler of a column, etc

Characteristics of Two
-Stage
Two-Stage
Stochastic Optimization Models
Second-Stage Decisions
Taken in order to adapt the plan or design to the uncertain parameters
realization.
Also called Recourse decisions.
Represented by Control Variables.
Example: the operating level; the production slate of a plant.
Sometimes first stage decisions can be treated as second stage decisions.
In such case the problem is called a multiple stage problem.

Example: Vinyl Chloride Plant

HCl recycle

Air or O2
Oxychlorination
Ethylene

Cl2

Light ends

EDC
purification
Direct
chlorination

EDC
pyrolysis

VCM
purification

VCM

EDC recycle
Heavy ends

Example: Vinyl Chloride Plant

Consider the following forecasts:
Forecasted prices of raw materials product
Year

Forecasted excess demand over current capacity

Ethylene

Chlorine

Oxygen

VCM

$/ton

$/ton

$/ft3

$/ton

2004

492.55

212.21

0.00144

499.19

2004

3602

2005

499.39

214.14

0.00144

506.19

2005

5521

2006

506.22

216.07

0.00143

513.18

2006

7355

2007

513.06

218.00

0.00142

520.18

2007

9551

2008

519.90

219.93

0.00141

527.18

2008

11888

2009

526.73

221.86

0.00140

529.17

2009

14322

2010

533.57

223.79

0.00139

535.17

2010

16535

2011

540.41

225.72

0.00138

543.17

2011

18972

Std. Dev

24.17

10.56

0.00010

26.15

Year

VCM
lb-mol/hr

Consider building (in 2004) for three capacities to satisfy excess demand at 2004,
2006 and 2011. Plants will operate under capacity until 2006 or 2011 in the last
two cases. These are 3 different first stage decisions.
6

Example: Vinyl Chloride Plant

The different investment costs are:
Plant Capacity

4090
MMlb/yr

6440
MMlb/yr

10500
MMlb/yr

TCI

$47,110,219

$68,886,317

$77,154,892

Consider the following calculation procedure

Random Number
Generation

Random numbers are

Raw material
Cost

Total Product
Cost

Income from
selling VCM

Gross Income

Cash Flow

Net Profit

obtained for each year for

raw materials and product
prices using sampling from a

Net Present
Worth

normal distribution. This can

be done in Excel

Risk &
Probability
7

Example: Vinyl Chloride Plant

Probability vs. Net Present Worth

Histograms and Risk Curves are

0.35
0.3
Probability

0.25

Notice the asymmetry in the distributions.

0.2
0.15
0.1
0.05
0
-4500

Risk at Different Capacity

-1500
0 6 1500
NPW ($10 )

6.44E9 lb/yr

Cummulative Probability

-3000

4.09E9 lb/yr

3000

1.05E10 lb/yr

0.8
0.6

The risk curves show a 36% chance of losing

0.4

money for the 10.5 billion lbs/year capacity,

0.2

31.7% for the 6.44 billion lbs/yr capacity and

0
-6.00E+09

-4.00E+09

-2.00E+09

0.00E+00

2.00E+09

4.00E+09

Net Present Worth ($)

6.44E9 lb/yr

4.09E9 lb/yr

6.00E+09

41% chance for the 4.09 billion lbs/year capacity.

Expected Profits are: 24%, 25% and 20%.

1.05E10 lb/yr

Capacity Planning
GIVEN:

Process Network

Forecasted Data

DETERMINE:

Set of Processes
Set of Chemicals

B
A

Demands & Availabilities

Costs & Prices
Capital Budget

Network Expansions

Timing
Sizing
Location

Production Levels

OBJECTIVES:

Maximize Net Present Value

Capacity Investment Planning

Design Variables: to be decided before the uncertainty reveals
x=

{Yit , Eit , Qit }

Y: Decision of building process i in period t

E: Capacity expansion of process i in period t
Q: Total capacity of process i in period t

Control Variables: selected after uncertain parameters become known.

Assume them known for the time being!!!!

ys =

{ S jlt , P jlt , Wit }

S: Sales of product j in market l at time t and scenario s
P: Purchase of raw mat. j in market l at time t and scenario s
W: Operating level of of process i in period t and scenario s

10

Example
Project Staged in 3 Time Periods of 2, 2.5, 3.5 years

Chemical 5

Chemical 1

Process 1

Process 2

Chemical 6

Process 5

Chemical 8

Chemical 2

Chemical 7

Process 3
Chemical 4

Chemical 3
Process 4

11

Example
One feasible (not necessarily optimal) solution could be
Chemical 5

Period 1
2 years

Process 1

Chemical 1

Process 2

Chemical 6

Chemical 2

Period 2
2.5 years

Chemical 5
Chemical 1

Process 1

Process 2

Chemical 6

Chemical 2
Chemical 7
Process 3
Chemical 3

Period 3
3.5 years

Process 5

Chemical 8

Chemical 4

Same flowsheet different production rates

12

MODEL
SETS

VARIABLES

I : Processes i,=1,,NP
J : Raw materials and Products, j=1,,NC
T: Time periods. T=1,,NT
L: Markets, l=1,..NM
Yit: An expansion of process I in period t takes place (Yit=1), does not take place (Yit=0)
Eit: Expansion of capacity of process i in period t.
Qit: Capacity of process i in period t.
Wit: Utilized capacity of process i in period t.
Pjlt : Amount of raw material/intermediate product j consumed from market l in period t
Sjlt : Amount of intermediate product/product j sold in market l in period t

ij : Amount of raw material/intermediate product j used by process i

ij : Amount of product/intermediate product j consumed by process i
jlt : Sale price of product/intermediate product j in market l in period t
jlt : Cost of product/intermediate product j in market l in period t
: Operating cost of process i in period t
PARAMETERS it
it : Variable cost of expansion for process i in period t
it : Fixed cost of expansion for process i in period t
Lt : Discount factor for period t
E itL , E itU :Lower and upper bounds on a process expansion in period t
a Ljlt , a Ujlt : Lower and upper bounds on availability of raw material j in market l in period t
d Ljlt , d Ujlt : Lower and upper bounds on demand of product j in market l in period t
CI t : Maximum capital available in period t

NEXPt: maximum number of expansions in period t

13

MATHEMATICAL PROGRAMMING MODEL

OBJECTIVE FUNCTION

Max NPV =

NT

t =1

NP
NP NT

NM NC
Lt ( jlt S jlt jlt Pjlt ) itWit ( it Eit + itYit )
i =1
i =1 t =1

l =1 j =1

DISCOUNTED REVENUES

INVESTMENT

Yit: An expansion of process I in period t takes place (Yit=1), does not take place (Yit=0)
Eit: Expansion of capacity of process i in period t.
Wit: Utilized capacity of process i in period t.
Pjlt : Amount of raw material/interm. product j consumed from market l in period t
Sjlt : Amount of intermediate product/product j sold in market l in period t
I : Processes i,=1,,NP
J : Raw mat./Products, j=1,,NC
T: Time periods. T=1,,NT
L: Markets, l=1,..NM

jlt : Sale price of product/intermediate product j in market l in period t

jlt : Cost of product/intermediate product j in market l in period t
it : Operating cost of process i in period t
it : Variable cost of expansion for process i in period t
it : Fixed cost of expansion for process i in period t
Lt : Discount factor for period t

14

MODEL
LIMITS ON EXPANSION

Yit EitL Eit Yit EitL

i =1,K, NP t =1,K, NT

TOTAL CAPACITY IN
EACH PERIOD

Qit = Qi ( t 1) + E it

i =1,K, NP t =1,K, NT

LIMIT ON THE NUMBER

OF EXPANSIONS
LIMIT ON THE CAPITAL
INVESTMENT

NT

Y
t =1

it

NP

(
i =1

it

NEXPi
E it + it Yit ) CI t

i =1,K, NP
t =1,K, NT

Yit: An expansion of process I in period t takes place (Yit=1), does not take place (Yit=0)
Eit: Expansion of capacity of process i in period t.
Qit: Capacity of process i in period t.

I : Processes i,=1,,NP
J : Raw mat./Products, j=1,,NC
T: Time periods. T=1,,NT
L: Markets, l=1,..NM

NEXPt: maximum number of expansions in period t

it : Variable cost of expansion for process i in period t
it : Fixed cost of expansion for process i in period t
E itL , E itU : Lower and upper bounds on a process expansion in period t
15

MODEL
UTILIZED CAPACITY IS
LOWER THAN TOTAL
CAPACITY

Wit Qit
NM

MATERIAL BALANCE

l =1

a Ljlt Pjlt a Ujlt

BOUNDS
NONNEGATIVITY

jlt

i =1,K, NP t =1,K, NT

NP

NM

NP

i =1

l =1

i =1

+ ijWit = S jlt + ijWit

d Ljlt S jlt d Ujlt

Eit , Qit , Wit , Pjlt , S jlt 0

i, j , l , t

i =1,K, NP t =1,K, NT

j = 1,..., NC , t = 1,..., NT , l = 1,..., NM

INTEGER
VARIABLES

Yit {0,1}
i =1,K, NP t =1,K, NT

Yit: An expansion of process I in period t takes place (Yit=1), does not take place (Yit=0)
Eit: Expansion of capacity of process i in period t.
Qit: Capacity of process i in period t.
Wit: Utilized capacity of process i in period t.
Pjlt : Amount of raw material/intermediate product j consumed from market l in period t
Sjlt : Amount of intermediate product/product j sold in market l in period t

I : Processes i,=1,,NP
J : Raw mat./Products, j=1,,NC
T: Time periods. T=1,,NT
L: Markets, l=1,..NM

a Ljlt , a Ujlt : Lower and upper bounds on availability of raw material j in market l in period t
d Ljlt , d Ujlt : Lower and upper bounds on demand of product j in market l in period t
16

MODEL
NM

MATERIAL BALANCE

l =1

jlt

NP

NM

NP

i =1

l =1

i =1

i =1,K, NP t =1,K, NT

+ ijWit S jlt + ijWit

A
i
D

NM

P
l =1

Blt

NM

+ kAWkt = S Clt + iDWit

l =1

Reference Component is C

kA , iD
Stoichiometric Coefficients

17

Two
-Stage Stochastic Formulation
Two-Stage
Let us leave it linear
because as is it is
complex enough.!!!

LINEAR MODEL SP

Max ps qsT ys cT x
s

Technology matrix

s.t.

First-Stage
Cost

Ax= b

First-Stage Constraints

Ts x +Wys = hs

Second-Stage Constraints

x0
First stage variables

Recourse
Function

x X

Second Stage Variables

ys 0
Recourse matrix (Fixed Recourse)
Sometimes not fixed (Interest rates in Portfolio Optimization)

Complete recourse: the

recourse cost (or profit) for
every possible uncertainty
realization remains finite,
independently of the first-stage
decisions (x).
Relatively complete recourse:
the recourse cost (or profit) is
feasible for the set of feasible
first-stage decisions. This
condition means that for every
feasible first-stage decision,
there is a way of adapting the
plan to the realization of
uncertain parameters.
We also have found that one
can sacrifice efficiency for
certain scenarios to improve
risk management. We do not
know how to call this yet.

18

Capacity Planning Under Uncertainty

GIVEN:

Process Network

Forecasted Data

DETERMINE:

Set of Processes
Set of Chemicals

B
A

Demands & Availabilities

Costs & Prices
Capital Budget

Network Expansions

Timing
Sizing
Location

Production Levels

OBJECTIVES:

Maximize Expected Net Present Value

Minimize Financial Risk
19

Process Planning Under Uncertainty

Design Variables: to be decided before the uncertainty reveals
x=

{Yit , Eit , Qit }

Y: Decision of building process i in period t

E: Capacity expansion of process i in period t
Q: Total capacity of process i in period t

Control Variables: selected after the uncertain parameters become known

ys =

{ Sjlts , Pjlts , Wits}

S: Sales of product j in market l at time t and scenario s

P: Purchase of raw mat. j in market l at time t and scenario s
W: Operating level of of process i in period t and scenario s

20

MODEL
LIMITS ON EXPANSION

Yit EitL Eit Yit EitL

i =1,K, NP t =1,K, NT

TOTAL CAPACITY IN
EACH PERIOD

Qit = Qi ( t 1) + E it

i =1,K, NP t =1,K, NT

LIMIT ON THE NUMBER

OF EXPANSIONS
LIMIT ON THE CAPITAL
INVESTMENT

NT

Y
t =1

it

NP

(
i =1

it

NEXPi
E it + it Yit ) CI t

i =1,K, NP
t =1,K, NT

Yit: An expansion of process I in period t takes place (Yit=1), does not take place (Yit=0)
Eit: Expansion of capacity of process i in period t.
Qit: Capacity of process i in period t.

I : Processes i,=1,,NP
J : Raw mat./Products, j=1,,NC
T: Time periods. T=1,,NT
L: Markets, l=1,..NM

NEXPt: maximum number of expansions in period t

it : Variable cost of expansion for process i in period t
it : Fixed cost of expansion for process i in period t
E itL , E itU : Lower and upper bounds on a process expansion in period t
21

MODEL
UTILIZED CAPACITY IS
LOWER THAN TOTAL
CAPACITY
NM

MATERIAL BALANCE

l =1

a Ljlts Pjlts aUjlts

BOUNDS
NONNEGATIVITY

Wits Qit

jlts

NP

NM

NP

i =1

l =1

i =1

+ ijWits S jlts + ijWits

d Ljlts S jlts d Ujlts

Eit , Qit ,Wits , Pjlts , S jlts 0

Yit {0,1}

i =1,K, NP t =1,K, NT

i, j , l , t

i =1,K, NP t =1,K, NT , s

j = 1,..., NC , t = 1,..., NT , l = 1,..., NM , s

INTEGER
VARIABLES

Yit: An expansion of process I in period t takes place (Yit=1), does not take place (Yit=0)
Eit: Expansion of capacity of process i in period t.
Qit: Capacity of process i in period t.
Wit: Utilized capacity of process i in period t.
Pjlt : Amount of raw material/intermediate product j consumed from market l in period t
Sjlt : Amount of intermediate product/product j sold in market l in period t

I : Processes i,=1,,NP
J : Raw mat./Products, j=1,,NC
T: Time periods. T=1,,NT
L: Markets, l=1,..NM

a Ljlts , aUjlts :Lower and upper bounds on availability of raw material j in market l in period t, scenario s
d Ljlts , d Ujlts : Lower and upper bounds on demand of product j in market l in period t, scenario s
22

MODEL
OBJECTIVE FUNCTION
NP
NT NM NC
NP NT
Max NPV = ps Lt ( jlts S jlts jlts Pjlts ) itsWits ( it Eit + itYit )
t =1 l =1 j =1
s
i =1
i =1 t =1

DISCOUNTED REVENUES

INVESTMENT

Yit: An expansion of process I in period t takes place (Yit=1), does not take place (Yit=0)
Eit: Expansion of capacity of process i in period t.
Wit: Utilized capacity of process i in period t.
Pjlt : Amount of raw material/interm. product j consumed from market l in period t
Sjlt : Amount of intermediate product/product j sold in market l in period t
I : Processes i,=1,,NP
J : Raw mat./Products, j=1,,NC
T: Time periods. T=1,,NT
L: Markets, l=1,..NM

jlt : Sale price of product/intermediate product j in market l in period t

jlt : Cost of product/intermediate product j in market l in period t
it : Operating cost of process i in period t
it : Variable cost of expansion for process i in period t
it : Fixed cost of expansion for process i in period t
Lt : Discount factor for period t

23

Example
Uncertain Parameters: Demands, Availabilities, Sales Price, Purchase Price
Total of 400 Scenarios
Project Staged in 3 Time Periods of 2, 2.5, 3.5 years
Chemical 5

Chemical 1

Process 1

Process 2

Chemical 6

Process 5

Chemical 8

Chemical 2

Chemical 7

Process 3
Chemical 4

Chemical 3
Process 4

24

Example Solution with Max ENPV

Period 1
2 years
Chemical 5

5.27 kton/yr
Chemical 1

5.27 kton/yr

Process 1

10.23 kton/yr
Chemical 7

19.60 kton/yr
Process 3
Chemical 3

22.73 kton/yr

19.60 kton/yr

25

Example Solution with Max ENPV

Period 2
2.5 years
Chemical 5

4.71 kton/yr
Chemical 1

4.71 kton/yr

Process 1

10.23 kton/yr

20.87 kton/yr
Process 3
Chemical 3

41.75 kton/yr

Chemical 7

22.73 kton/yr

Process 5

22.73 kton/yr

Chemical 8

20.87 kton/yr

Process 4

22.73 kton/yr

Chemical 4
20.87 kton/yr

26

Example Solution with Max ENPV

Period 3
3.5 years

14.95 kton/yr
Chemical 5

29.49 kton/yr
Chemical 1

44.44 kton/yr

Process 2

Process 1

Chemical 6

29.49 kton/yr
80.77 kton/yr

80.77 kton/yr
Chemical 2
29.49 kton/yr

21.88 kton/yr
Chemical 7
Process 5

Process 3
Chemical 3

43.77 kton/yr

Chemical 8

21.88 kton/yr

22.73 kton/yr

22.73 ton/yr

Process 4

22.73 kton/yr

Chemical 4
21.88 kton/yr

27

Example Solution with Max ENPV

Risk
1.0

PP solution

0.9
0.8
0.7
0.6
0.5
0.4

E[NPV ] = 1140 M$

0.3
0.2
0.1
0.0
250

500

750

1000

1250

1500

1750

2000

2250

2500

2750

3000

3250

NPV (M$)

28

Example Risk Management Solutions

Risk
1.0
0.9
0.8

increases

0.7

PP

0.6

500
600

0.5

700
800

0.4

900
1000

0.3

1100
1200

0.2

1300
1400

0.1

1500

0.0
250

500

750

1000

1250

1500

1750

2000

2250

2500

2750

3000

3250

NPV (M$)

29

Example Risk Management Solutions

Risk
1.0
0.9

= 900
= 1100

ENPV = 908

0.8

PP
ENPV =1140

ENPV = 1074
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
250

500

750

1000

1250

1500

1750

2000

2250

2500

2750

3000

3250

NPV (M$)

30

Example Risk Management Solutions

NPV PDF f ()
0.0026
0.0024
0.0022

= 900

0.0020
0.0018
0.0016
0.0014
0.0012
0.0010
0.0008

= 1100

PP

0.0006
0.0004
0.0002
0.0000
0

500

1000

1500

2000

2500

3000

NPV ( , M$ )

31

Example Solution with Min DRisk

(=900)
DRisk(=900)
Period 1
2 years

- Same Flowsheet
- Different Capacities

Chemical 5
5.59 kton/yr
Chemical 1
5.59 kton/yr

Process 1
10.85 kton/yr
Chemical 7
19.30 kton/yr
Process 3

Chemical 3

22.37 kton/yr

19.30 kton/yr

32

Example Solution with Min DRisk

(=900)
DRisk(=900)
Period 2
2.5 years

- Same Flowsheet
- Different Capacities

Chemical 5
4.99 kton/yr
Chemical 1
4.99 kton/yr

Process 1
10.85 kton/yr

20.85 kton/yr
Process 3
Chemical 3

Chemical 7

22.37 kton/yr

Process 5
22.43 kton/yr

Chemical 8
20.85 kton/yr

41.70 kton/yr
Process 4
22.37 kton/yr

Chemical 4
20.85 kton/yr

33

Example Solution with Min DRisk

(=900)
DRisk(=900)
Period 3
3.5 years

- Same Flowsheet
- Different Capacities
- No Expansion of Process 1
2.39 kton/yr
Chemical 5
5.15 kton/yr
Chemical 1
7.54 kton/yr

Process 1

Process 2

Chemical 6
5.15 kton/yr

10.85 kton/yr

10.85 kton/yr
Chemical 2
5.15 kton/yr

21.77 kton/yr
Chemical 7
Process 5

Process 3
Chemical 3
43.54 kton/yr

Chemical 8
21.77 kton/yr

22.37 kton/yr

22.77 ton/yr

Process 4
22.37 kton/yr

Chemical 4
21.77 kton/yr

34

RECENT RESULTS
Gas Commercialization in Asia
Network of Alternatives
Suppliers
Australia
Indonesia
Iran
Kazakhstan
Malaysia
Qatar
Russia

Transportation
Methods
Pipeline

Markets
China

LNG

India

CNG

Japan

GTL

S. Korea

Ammonia

Thailand

Methanol

United States
35

Gas Commercialization in Asia

Some solutions
Processing Facilities

1.0

OV @ 95%: 1.42

Time
Period
T1
T2
T3
T4
T5
T6

OV @ 95%: 1.75

0.9
0.8
0.7
0.6

Mala-GTL
Ships: 4 & 2
ENPV:4.570
DR@ 4: 0.157
DR@ 3.5: 0.058

0.5
0.4
0.3

Indo-GTL
Ships: 5 & 3
ENPV:4.633
DR@ 4: 0.190
DR@ 3.5: 0.086

0.2
0.1

VaR @ 5%: 1.82

1

VaR @ 5%: 1.49

3

FCI
3.00
0.00
1.89
0.00
0.00
0.00

Cap Flow Feed

0.00
4.57
4.57
7.49
7.49
7.49

0.00
4.47
4.57
7.32
7.49
7.49

0.0
297.9
304.9
488.2
499.6
499.6

Transportation to:
China

10

Time
Period
T1
T2
T3
T4
T5
T6

Flow

Ships

Flow

Avrg.
Ships

0.00
1.16
0.00
0.42
0.00
0.00

0.00
0.98
0.00
0.35
0.00
0.00

0.00
2.79
3.66
5.58
6.00
6.00

0.00
3.49
4.57
6.97
7.49
7.49

0.00
3.95
3.66
6.00
6.00
6.00

4.0
4.0
6.0
6.0
6.0

Indo (GTL)

FCI
3.00
0.00
1.90
0.00
0.00
0.00

Cap Flow Feed

0.00
4.43
4.43
7.18
7.18
7.18

0.00
4.25
4.43
7.09
7.18
7.18

0.0
283.1
295.5
472.6
479.0
479.0

Thai

Ships Ships

Processing Facilities

0.0
0

Mala (GTL)

Transportation to:
China

Thai

Ships Ships

Flow

Ships

Flow

Avrg.
Ships

0.00
1.12
0.00
0.44
0.00
0.00

0.00
0.76
0.00
0.30
0.00
0.00

0.00
3.88
4.94
7.56
8.00
8.00

0.00
3.48
4.43
6.79
7.18
7.18

0.00
5.00
4.94
8.00
8.00
8.00

5.0
5.0
8.0
8.0
8.0

36

Conclusions
Risk is usually assessed after a design/plan has been selected but it cannot be
managed during the optimization stage (even when stochastic optimization
including uncertainty has been performed).
If Risk is to be managed, the decision maker has two simultaneous objectives:

Maximize Expected Profit.

Minimize Risk Exposure

There are some good ways of obtaining good solutions (next lecture).

37