You are on page 1of 31

Introduction and Motivation

Closely Related Work


Methodology
Example Illustration
Summary

Fenchel Disjunctive Decomposition for Mean-Risk


Stochastic Integer Programs

Lewis Ntaimo
Michelle M. Alvarado and Guglielmo Lulli†

Industrial & Systems Engineering Department, Texas A&M University, College


Station, Texas, USA
† Dipartimentodi Informatica Sistemistica e Comunicazione
Universitá degli Studi di Milano–Bicocca
Milan, Italy

ICSP 2013 Bergamo, Italy

1 / 31
Introduction and Motivation
Closely Related Work
Methodology
Example Illustration
Summary

Agenda

Introduction and Motivation

Closely Related Work

Methodology
Subgradient Decomposition
Fenchel Decomposition (FD)
Fenchel Disjunctive Decomposition (FDD)

Example Illustration

Summary

2 / 31
Introduction and Motivation
Closely Related Work
Methodology
Example Illustration
Summary

Two-stage stochastic mixed-integer program (SMIP) with mean risk objective:

SIP2: Min E[f (x, ω̃)] + λD[f (x, ω̃)] (1)


x∈X

where for an outcome ω of ω̃:


Q(x, ω) = Min q > y (ω)
s.t. Wy (ω) ≥ h(ω) − T (ω)x (2)
n0
y (ω) ∈ Zn̄+2 × R+2 ,

n0
 x ∈ Zn̄+1 × R+1 is a decision variable vector
 λ > 0 is a suitable weight factor
n0
 X ⊆ Rm1 is a nonempty polyhedron, X = {Ax ≥ b, x ∈ Zn̄+1 × R+1 }
 ω̃ multivariate random variable and for a given x ∈ X the real random cost
variable is given as: f (x, ω̃) = c > x + Q(x, ω̃)

3 / 31
Introduction and Motivation
Closely Related Work
Methodology
Example Illustration
Summary

More Notation:
 The family of real random cost variables {f (x, ω̃)}x∈X ⊆ F are defined on a
probability space (Ω, F, P)
 E : F 7→ R denotes the expected value
 D : F 7→ R is a mean-risk measure
 F is the space of all real random cost variables f : Ω 7→ R
 A scenario defines the realization of the stochastic problem data
{h(ω), T (ω)}
I W ∈ Rm2 ×n2 is a fixed recourse matrix
Assumptions:
(A1) The random variable ω̃ is discrete with finitely many scenarios ω ∈ Ω, each
with probability of occurrence p(ω).
(A2) The first-stage feasible set is nonempty, that is, X 6= ∅.
n0
(A3) The second-stage feasible set {Wy (ω) ≥ h(ω) − T (ω)x, y (ω) ∈ Zn̄+2 × R+2 }
is bounded and nonempty for all x ∈ X .

4 / 31
Introduction and Motivation
Closely Related Work
Methodology
Example Illustration
Summary

Mean-Risk Measures
Expected Excess (EE): EE is the expected value of the excess over the target
η ∈ R.
D = φEEη (x) = E[max{f (x, ω̃) − η, 0}].
Absolute Semideviation (ASD): ASD reflects the expected value of the excess
over the mean value.

D = φASD (x) = E[max{f (x, ω̃) − E[f (x, ω̃)], 0}].

ASD Deterministic Equivalent Form: Given λ ∈ [0, 1]


X X
ASD2: Min (1 − λ)c > x + (1 − λ) p(ω)q > y (ω) + λ p(ω)ν(ω) (3a)
ω∈Ω ω∈Ω
s.t. T (ω)x + Wy (ω) ≥ h(ω), ∀ω ∈ Ω
− c > x − q > y (ω) + ν(ω) ≥ 0, ∀ω ∈ Ω
X
−c > x − p(ω)q > y (ω) + ν(ω) ≥ 0, ∀ω ∈ Ω (3b)
ω∈Ω
n0
x ∈ X , y (ω) ∈ Zn̄+2 × R+2 , ν(ω) ∈ R, ∀ω ∈ Ω.

5 / 31
Introduction and Motivation
Closely Related Work
Methodology
Example Illustration
Summary

Motivation:
Limited algorithms available to solve this class of problems despite many
applications
 Example approach: Dual decomposition based (Schultz, 2011)
 Applications of interest: chemotherapy scheduling, wildfire risk
management, air traffic flow management
Properties and challenges:
 Large-scale nature, mean-risk objective function is lower
semi-continuous, generally nonconvex and possibly discontinuous
(Schultz, 2011)
 Decomposing mean-risk SMIP based on semideviation scenario-wise
does not result in separable scenario problems

6 / 31
Introduction and Motivation
Closely Related Work
Methodology
Example Illustration
Summary

Related Work
SMIP with Mean-Risk Description Decision Vars.
(Märkert and Schultz, 2005) Deviation measures mixed-integer
(Schultz and Tiedemann, 2006) CVAR mixed-integer
(Miller and Ruszczyn’ski, 2009) QDEV continuous
(Schultz, 2011) Excess prob., CVAR, EE, mixed-integer
and ASD
Relevant Methodology
(Carøe, 1998) Disjunctive cuts integer
(Carøe and Schultz, 1999) Dual decomposition integer
(Sen and Higle, 2005) Disjunctive decomposition mixed-binary
(Ntaimo, 2013) Fenchel decomposition (FD) mixed-integer
(Ahmed, 2006) Subgradient decomposition continuous
(Cotton and Ntaimo, 2013) Subgradient decomposition continuous

7 / 31
Introduction and Motivation
Closely Related Work Subgradient Decomposition
Methodology Fenchel Decomposition (FD)
Example Illustration Fenchel Disjunctive Decomposition (FDD)
Summary

Methodology
 Subgradient Decomposition
I Stage-wise decomposition: x, y (ω)
I Generate subgradients to sequentially
approximate convex E and D terms

 Fenchel Decomposition
I Generate FD cuts to cut off fractional
LP-relaxation solution
I FD cuts are in y (ω)-space or (x, y (ω))-space
I Use disjunctive programming to lift cuts to
(x, y (ω))-space

 Fenchel Disjunctive Decomposition


I Fixed recourse, maintain FD cuts with
common-cut-coefficients in y (ω)-space
I Lift/translate the cuts to the (x, y (ω))-space
for each ω ∈ Ω
8 / 31
Introduction and Motivation
Closely Related Work Subgradient Decomposition
Methodology Fenchel Decomposition (FD)
Example Illustration Fenchel Disjunctive Decomposition (FDD)
Summary

Subgradient Decomposition

Master Problem:
Min (1 − λ)c > x k + (1 − λ)θ + λζ
s.t. ψτ> x + θ ≥ ψτ0 , τ = 1, · · · , k
(4)
στ> x + ζ ≥ στ0 , τ = 1, · · · , k
x ∈ X , θ, ζ free

Subproblem Relaxation: for ω ∈ Ω

Q(x k , ω) = Min q > y (ω)


s.t. Wy (ω) ≥ h(ω) − T (ω)x k
−βt> y (ω) ≥ −γt (ω) + αt (ω)x k ,
t = 1, · · · , k
y (ω) ≥ 0
(5)
9 / 31
Introduction and Motivation
Closely Related Work Subgradient Decomposition
Methodology Fenchel Decomposition (FD)
Example Illustration Fenchel Disjunctive Decomposition (FDD)
Summary

Subgradient Decomposition

Master Problem: Computational study of subgradient


> k decomposition for mean-risk SLP
Min (1 − λ)c x + (1 − λ)θ + λζ
(Cotton and Ntaimo, 2013)
s.t. ψτ> x + θ ≥ ψτ0 , τ = 1, · · · , k
(4)  Separate cuts (Ahmed, 2006)
στ> x + ζ ≥ στ0 , τ = 1, · · · , k versus aggregated cuts
x ∈ X , θ, ζ free  Quantile and deviation risk
measures
Subproblem Relaxation: for ω ∈ Ω  computations with standard
instances:
Q(x k , ω) = Min q > y (ω)
I mean-risk approach
s.t. Wy (ω) ≥ h(ω) − T (ω)x k
appropriate in non-uniform
−βt> y (ω) ≥ −γt (ω) + αt (ω)x k , distribution cases
t = 1, · · · , k I ASD more conservative than
y (ω) ≥ 0 QDEV
(5)
10 / 31
algorithm is superior. However, for QDEV instance clearly QDEV-SEP is the
Introduction and Motivation
preferred algorithm. It performs
Closely betterSubgradient
Related Work in termsDecomposition
of both algorithm iterations
Methodology Fenchel Decomposition (FD)
and CPU time. Example Illustration Fenchel Disjunctive Decomposition (FDD)
Summary

Sample results: Aggregate (AGG) and separate (SEP) cut algorithms


Table 3 Performance of ASD-AGG and ASD-SEP algorithms
ASD-AGG ASD-SEP
Iters CPU(s) Iters CPU(s)
Instance Avg Stdev Avg Stdev Avg Stdev Avg Stdev
cep1 2.00 0.00 0.02 0.01 2.00 0.00 0.02 0.00
cep1a 2.00 0.00 0.03 0.01 2.00 0.00 0.02 0.00
cep1sk 6.00 0.00 0.04 0.03 6.00 0.00 0.05 0.02
pgp2 33.91 2.63 0.57 0.08 31.45 2.30 0.55 0.09
pgp2e 37.91 2.63 0.52 0.07 35.36 2.50 0.51 0.06
pgp2f 37.64 3.29 0.52 0.10 33.09 3.83 0.45 0.07
gbd 33.55 3.72 258.18 30.52 31.55 2.02 242.43 24.18
gbd sk3 27.09 7.37 186.67 51.26 22.73 5.00 154.02 33.74
LandS 31.00 1.61 389.40 18.61 30.18 1.25 382.81 32.09
20Tr l 2250.91 332.88 1227.78 217.13 2185.73 572.10 1350.58 507.60
20Tr lsk1 1738.00 152.01 972.01 108.19 1587.00 166.80 914.05 112.60
20Tr m 1912.09 316.13 1784.00 318.54 1804.36 233.99 1739.86 244.41
20Tr msk1 2108.00 217.13 2310.71 287.26 1783.18 246.02 1995.83 382.34
20Tr msk2 1865.55 238.40 1861.99 297.92 1676.91 238.75 1738.45 321.48
20Tr h 1757.00 106.73 3088.27 218.56 1505.45 165.88 2631.08 273.72
ssnTr l 352.91 149.69 90.51 38.57 317.09 151.55 86.45 41.47
ssnTr m 983.45 1339.86 1647.08 2318.09 890.73 990.84 1491.98 1699.64
ssnTr h 788.40 300.45 6929.49 2620.53 1013.55 423.16 8876.52 3665.14
stormTr l 16.00 0.00 2.36 0.09 16.00 0.00 2.42 0.09
stormTr m 13.00 0.00 47.68 0.35 13.00 0.00 48.25 0.47
stormTr h 13.00 0.00 1193.81 5.32 13.00 0.00 1195.29 10.24

11 / 31
Introduction and Motivation
Closely Related Work Subgradient Decomposition
Methodology Fenchel Decomposition (FD)
Example Illustration Fenchel Disjunctive Decomposition (FDD)
Summary

Fenchel Decomposition (FD)

 Risk neutral case


 Exploit special structure:
 Random or fixed W , mixed-binary decision variables
 Special constraints: e.g. knapsack constraints
 For each scenario ω, either
 Generate cuts of the form α> x + β(ω)> y ≤ g (ω, β)
or
 Generate cuts of the form β(ω)> y ≤ g (ω, x, β(ω)) and use disjunctive
programming to lift the cut β(ω)> y ≤ g (ω, x̂, β(ω)) to the
(x, y (ω))-space so that it is valid for all x ∈ X

12 / 31
Introduction and Motivation
Closely Related Work Subgradient Decomposition
Methodology Fenchel Decomposition (FD)
Example Illustration Fenchel Disjunctive Decomposition (FDD)
Summary

For (ω, x) ∈ Ω × X define sets for y (ω)-space:


 YLP (ω, x) = y ∈ Rn+2 | Wy ≥


h(ω) − T (ω)x

 YIP (ω, x) = y ∈ YLP (ω, x) | y ∈
n0
Bn̄+2 × R+2

c (ω, x)
 Convex hull of YIP (ω, x): YIP

For (ω) ∈ Ω define sets for (x, y (ω))-space:


n0
 YLP (ω) = x ∈ Bn̄+1 × R+1 , y ∈ Rn+2 |


Ax ≥ b, T (ω)x + Wy ≥ h(ω)

 YIP (ω) = (x, y ) ∈ YLP (ω) | y ∈
n0
Bn̄+2 × R+2

c (ω)
 Convex hull of YIP (ω): YIP

13 / 31
Introduction and Motivation
Closely Related Work Subgradient Decomposition
Methodology Fenchel Decomposition (FD)
Example Illustration Fenchel Disjunctive Decomposition (FDD)
Summary

Key result for generating FD cuts:

Theorem
Let ω̄ denote a scenario with a fractional
subproblem LP-relaxation solution ŷ (ω̄) for a
given x̂ ∈ X . Define
g (ω̄, x̂, β) = Max {β > y | y ∈ YIP c (ω̄, x̂)} and

let δ(ω̄, x̂, β) = β > ŷ (ω̄) − g (ω̄, x̂, β). Then


there exists a vector β for which δ(ω̄, x̂, β) > 0
if and only if ŷ (ω̄) ∈
/ YIP c (ω̄, x̂).

I Theorem 1 allows for generating valid


c (ω̄, x̂). However, we
inequalities for YIP
are interested in valid inequalities for
c (ω, x) for all ω ∈ Ω.
YIP

14 / 31
Introduction and Motivation
Closely Related Work Subgradient Decomposition
Methodology Fenchel Decomposition (FD)
Example Illustration Fenchel Disjunctive Decomposition (FDD)
Summary

15 / 31
Introduction and Motivation
Closely Related Work Subgradient Decomposition
Methodology Fenchel Decomposition (FD)
Author's personal
Example copy
Illustration Fenchel Disjunctive Decomposition (FDD)
Author's personal copy
Summary
160 J Glob Optim (2013) 55:141–163
Sample results for risk neutral case: FD and FD-L algorithms
160 J Glob Optim (2013) 55:141–163
Table 9 Computational Results on SSLP using FD
Table 9 ComputationalFD
Instance Results on SSLP using FD
Algorithm CPU (s)
Instance FD
FD Algorithm
Iters FD cuts L2 cuts CPU
Min (s) Max Avg

SSLP5.25.50 FD2.80
Iters FD cuts
3.40 L2 cuts
0.00 Min0.37 Max0.69 Avg0.54

SSLP5.25.100
SSLP5.25.50 6.40
2.80 15.00
3.40 0.20
0.00 0.52
0.37 3.19
0.69 1.24
0.54
SSLP10.50.50
SSLP5.25.100 30.60
6.40 495.60
15.00 1.00
0.20 40.31
0.52 79.02
3.19 59.50
1.24
SSLP10.50.100
SSLP10.50.50 32.40
30.60 936.60
495.60 1.00
1.00 78.39
40.31 134.47
79.02 102.75
59.50
SSLP10.50.500
SSLP10.50.100 28.40
32.40 4669.60
936.60 1.00
1.00 376.54
78.39 517.81
134.47 451.80
102.75
SSLP15.45.5
SSLP10.50.500 94.60
28.40 321.20
4669.60 1.00
1.00 7.99
376.54 526.85
517.81 136.67
451.80
SSLP15.45.10
SSLP15.45.5 77.75
94.60 588.00
321.20 1.00
1.00 17.58
7.99 376.75
526.85 171.45
136.67
SSLP15.45.15
SSLP15.45.10 118.20
77.75 1318.40
588.00 1.00
1.00 22.61
17.58 771.80
376.75 338.37
171.45
SSLP15.45.20
SSLP15.45.15 78.40
118.20 1139.20
1318.40 1.00
1.00 100.06
22.61 300.25
771.80 166.72
338.37
SSLP15.45.25
SSLP15.45.20 83.50
78.40 1393.00
1139.20 1.00
1.00 85.41
100.06 244.59
300.25 168.69
166.72
SSLP15.45.25 83.50 1393.00 1.00 85.41 244.59 168.69
Table 10 Computational results on SSLP using FD-L
Table 10 Computational
Instance FD-Lresults on SSLP using FD-L
Algorithm CPU (s)
Instance FD-L
FD-L Algorithm
Iters FD-L cuts L2 cuts CPU
Min (s) Max Avg
FD-L Iters FD-L cuts L2 cuts Min Max Avg
SSLP5.25.50 2.40 2.80 0.00 0.36 0.70 0.52
SSLP5.25.100
SSLP5.25.50 6.00
2.40 12.80
2.80 0.00
0.00 0.37
0.36 2.56
0.70 1.10
0.52
SSLP10.50.50
SSLP5.25.100 31.40
6.00 503.20
12.80 1.00
0.00 37.51
0.37 75.79
2.56 57.74
1.10
SSLP10.50.50
SSLP10.50.100 31.40
31.20 503.20
899.40 1.00
1.00 37.51
69.44 75.79
108.48 57.74
91.57
SSLP10.50.100
SSLP10.50.500 31.20
28.60 899.40
4601.60 1.00
1.00 69.44
345.20 108.48
476.56 91.57
416.07
SSLP10.50.500
SSLP15.45.5 28.60
94.80 4601.60
327.20 1.00
1.00 345.20
7.83 476.56
553.60 416.07
138.61
SSLP15.45.5
SSLP15.45.10 94.80
84.25 327.20
637.50 1.00
1.00 7.83
27.39 553.60
404.66 138.61
181.53
SSLP15.45.10
SSLP15.45.15 84.25
117.00 637.50
1306.60 1.00
1.00 27.39
21.58 404.66
723.06 181.53
333.20
SSLP15.45.15
SSLP15.45.20 117.00
80.60 1306.60
1164.00 1.00
1.00 21.58
102.71 723.06
247.63 333.20
166.69
SSLP15.45.20
SSLP15.45.25 80.60
83.25 1164.00
1383.00 1.00
1.00 102.71
82.57 247.63
201.81 166.69
151.61
SSLP15.45.25 83.25 1383.00 1.00 82.57 201.81 151.61

on the first five problems. However, FD and FD-L perform better on the last five problems,
on the first
which havefive problems.
larger However,
first-stage FDdimension
decision and FD-Lspace.
perform betterthe
Unlike onFD
the and
last D2
five algorithms,
problems, 16 / 31
Introduction and Motivation
Closely Related Work Subgradient Decomposition
Methodology Fenchel Decomposition (FD)
Example Illustration Fenchel Disjunctive Decomposition (FDD)
Summary

Fenchel Disjunctive Decomposition (FDD)

 Fixed W , pure binary first-stage, mixed-binary second-stage


 Use FD to generate common-cut-coefficients:
 generate cut β > y ≤ g (ω̄, x̂, β) for ‘representative’ scenario ω̄
 Can select a scenario ω̄ ∈ Ω with a fractional solution ŷ (ω̄) or use a
‘conditional expected’ scenario ω̄, based on scenarios with LP fractional
solutions
 Use disjunctive programming:
 Lift and translate the cut β > y ≤ g (ω̄, x̂, β) to the (x, y (ω))-space for
each ω ∈ Ω so that it is valid for all x ∈ X

17 / 31
Introduction and Motivation
Closely Related Work Subgradient Decomposition
Methodology Fenchel Decomposition (FD)
Example Illustration Fenchel Disjunctive Decomposition (FDD)
Summary

Proposition
Suppose we have binary x̂ ∈ X . Then
c
conv {y | (x̂, y ) ∈ YIP (ω)} = {y | (x̂, y ) ∈ YIP (ω)}

(Carøe, 1998)

c
 This result relates facets of {y | (x̂, y ) ∈ YIP (ω)} to
conv {y | (x̂, y (ω)) ∈ YIP (ω)} for binary x̂ ∈ X .
 This implies that we can lift facets (valid inequalities) of the form
β > y ≤ g (ω, x̂, β) into the (x, y (ω))-space.
 To accomplish this, we use result by Araoz et al. (1984) on next slide
applied to our setting

18 / 31
Introduction and Motivation
Closely Related Work Subgradient Decomposition
Methodology Fenchel Decomposition (FD)
Example Illustration Fenchel Disjunctive Decomposition (FDD)
Summary

Proposition
The inequality α(ω̄)> x + ϕβ > y ≤ ϕg (ω̄, x̂, β) + α(ω̄)> x̂, where ϕ ≥ 0, is a facet
extension of β > y ≤ g (ω̄, x̂, β) if and only if ϕ > 0 and α(ω̄)/ϕ is a vertex of the
polyhedron

{α(ω̄) | α(ω̄)> x + β > y ≤ g (ω̄, x̂, β) is valid for YIP


c
(ω̄)} (6)

of valid inequalities; or ϕ = 0 and α(ω̄) is an extreme ray of (6).

 This result allows for β > y ≤ g (ω̄, x̂, β) to be lifted to a valid inequality for
c
YIP (ω̄) if the extreme points of (6) are known.
c
 We want valid inequalities for YIP (ω) for each ω ∈ Ω.

19 / 31
Introduction and Motivation
Closely Related Work Subgradient Decomposition
Methodology Fenchel Decomposition (FD)
Example Illustration Fenchel Disjunctive Decomposition (FDD)
Summary

Disjunctive programming:
Let i be an index for a binary component of z = (x, y ) and define
 R = {0, 1}.
 S0 (ω) = YLP (ω) ∩ {−zi ≥ 0} and S1 (ω) = YLP (ω) ∩ {zi ≥ 1}, where
Sr (ω) 6= ∅ for all r ∈ R.
 disjunction of the two sets: S(ω) = ∪r ∈R Sr (ω)

c c
 Since YIP (ω) ⊆ S(ω), valid inequalities for S(ω) are also valid for YIP (ω)
c
 So our aim is to derive valid inequalities for YIP (ω) via S(ω) using
disjunctive programming
I Motivation: Generating valid inequalities for S(ω) involves solving LPs
instead of a subgradient optimization problem

20 / 31
Introduction and Motivation
Closely Related Work Subgradient Decomposition
Methodology Fenchel Decomposition (FD)
Example Illustration Fenchel Disjunctive Decomposition (FDD)
Summary

Definition
The reverse polar of the set S(ω), denoted S ] (ω), is defined as follows:
S ] (ω) = {(α(ω), β, γ(ω)) | α(ω)> x + β > y ≤ γ(ω) is valid for S(ω)}.

I The set S ] (ω) defines the feasible set we need in order to compute α(ω)
and γ(ω) for each ω ∈ Ω given β derived based on ω̄
I The following result guarantees the existence of such α(ω) and γ(ω)

Theorem
Let W be fixed for all ω ∈ Ω. Let x̂ ∈ X (binary) and ŷ ∈ YLP (ω̄, x̂) be given and
suppose that β > y ≤ g (ω̄, x̂, β) is valid for YIP
c
(ω̄, x̂). Then there exists a vector
α(ω̄) ∈ Rn1 and scalar γ(ω̄) ∈ R such that α(ω̄)> x + β > y ≤ γ(ω̄) is valid for
c
YIP (ω̄)

I The proof of this theorem provides a way for computing α(ω̄) and γ(ω̄),
restricted to some convex set Πx 0 .

21 / 31
Introduction and Motivation
Closely Related Work Subgradient Decomposition
Methodology Fenchel Decomposition (FD)
Example Illustration Fenchel Disjunctive Decomposition (FDD)
Summary

 Given: (x̂, ŷ (ω̄)) ∈ X × YLP (ω, x̂) and β > y ≤ g (ω̄, x̂, β) that is valid for
c
YIP (ω̄, x̂)
 Generate: α(ω̄)> x + β > y ≤ γ(ω̄) valid for YIP
c
(ω̄) by solving this LP

22 / 31
Introduction and Motivation
Closely Related Work Subgradient Decomposition
Methodology Fenchel Decomposition (FD)
Example Illustration Fenchel Disjunctive Decomposition (FDD)
Summary

 Given: (x̂, ŷ (ω̄)) ∈ X × YLP (ω, x̂) and β > y ≤ g (ω̄, x̂, β) that is valid for
c
YIP (ω̄, x̂)
 Generate: α(ω̄)> x + β > y ≤ γ(ω̄) valid for YIP
c
(ω̄) by solving this LP

δ(ω̄) = Max α(ω̄)> x̂ − γ(ω̄)


(α(ω̄),γ(ω̄))∈Πx 0
> >
s.t. − α(ω̄) − u01 A − u02 T (ω̄) + u03 ei ≥ 0
> >
− α(ω̄) − u11 A − u12 T (ω̄) − u13 ei ≥ 0
>
− u02 W ≥β (7)
>
− u12 W ≥β
> >
u01 b + u02 h(ω̄) + γ(ω̄) ≥ 0
> >
u11 b + u12 h(ω̄) + u13 + γ(ω̄) ≥ 0
u01 , u02 , u03 , u11 , u12 , u13 ≥ 0

I Problem (7) characterizes the reverse polar S ] (ω̄) based on the disjunction
c
{−xi ≥ 0} ∪ {xi ≥ 1} and YIP (ω̄)

23 / 31
Introduction and Motivation
Closely Related Work Subgradient Decomposition
Methodology Fenchel Decomposition (FD)
Example Illustration Fenchel Disjunctive Decomposition (FDD)
Summary

FDD Algorithm
Step 0. Initialization.
Step 1. Solve Subproblem LPs.
I For each ω ∈ Ω solve subproblem (5). Step 3. Re-Solve Subproblem LPs
I If (x k , {y k (ω)}ω∈Ω ) satisfy integer I For each ω ∈ Ω solve the subproblem.
restrictions, compute optimality cuts for I Compute optimality cuts for E and D
E and D terms, update upper bound and
terms.
go to step 4. I If (x k , {y k (ω)}ω∈Ω ) satisfy integer
I Otherwise, go to step 2.
restrictions, update upper bound and go
Step 2. Generate Cutting Planes and Refine Feasible to step 4.
Region. I Otherwise, go to step 5.
I FD Cut: generate βk . Step 4. Termination.
I Form and solve separation problem to I Check termination criterion.
obtain βk and g (ω̄, αk (ω̄), βk ).
Step 5. Update and Solve the Master Problem.
I Form and solve reverse polar LP (7) to I Add optimality cuts to master problem
compute αk (ω) and γk (ω) for each (4).
ω ∈ Ω. I Solve master problem and update lower
I For each ω ∈ Ω, use βk from Step 2(i) bound.
to form problem (7) and solve to get I Repeat from Step 1.
αk (ω) and γk (ω).
I Form cut −βk> y ≥ −γk (ω) + αk (ω)> x k
and append to subproblem (5).

24 / 31
Introduction and Motivation
Closely Related Work
Methodology
Example Illustration
Summary

Numerical Example
Instance Data:
   
c = 2; A = −1 ; b = −1
Problem:  
60
q= ; p(ω1 ) = p(ω2 ) = 0.5;
20
Min E[f (x, ω̃)] + λD[f (x, ω̃)]  4 
3
−1
s.t. Ax ≥ b − 5 −1
x ∈ {0, 1} W =  −1
2 
0 
0 −1
where f (x, ω̃) = c > x + Q(x, ω̃) and
− 43 − 13
   
 −2   −2 
Q(x, ω) =Min q > y h(ω1 ) =  −1 ; h(ω2 ) =  −1 
  
s.t. Wy ≥ h(ω) − T (ω)x −1 −1
y ∈ {0, 1}2    
0 0
 2   −1 
D : Absolute semi-deviation (ASD) T (ω1 ) =  0 ; T (ω2 ) =  0 
  

0 0
Consider λ = 0.5

25 / 31
Introduction and Motivation
Closely Related Work
Methodology
Example Illustration
Summary

DEP:

z = Min x − 15y1 (ω1 ) + 5y2 (ω1 ) − 15y1 (ω2 ) + 5y2 (ω2 )


+ 0.25ν1 + 0.25ν2 Optimal Solution:
s.t. − x ≥ −1 z = −20.5
4 4 x =1
y1 (ω1 ) − y2 (ω1 ) ≥ − y (ω1 ) = (1, 0), y (ω2 ) = (0, 0)
3 3 ν1 = −28, ν2 = 2.
5
− y1 (ω1 ) − y2 (ω1 ) + 2x ≥ −2
2
− y1 (ω1 ) ≥ −1 LP solution:
− y2 (ω1 ) ≥ −1 zLP = −48.25
x = 0.25
4 1
y1 (ω2 ) − y2 (ω2 ) ≥ − y (ω1 ) = (1, 0), y (ω2 ) = (0.7, 0)
3 3 ν1 = −50.5, ν2 = −41.5.
5
− y1 (ω2 ) − y2 (ω2 ) − x ≥ −2
2 *******
− y1 (ω2 ) ≥ −1 FDD algorithm terminates with optimal
− y2 (ω2 ) ≥ −1 solution at k = 3:
60y1 (ω1 ) − 20y2 (ω1 ) − 2x + ν1 ≥ 0 z = −20.5, x = 1, y (ω1 ) =
(1, 0), y (ω2 ) = (0, 0) ν1 = −28, ν2 = 2
60y1 (ω2 ) − 20y2 (ω2 ) − 2x + ν2 ≥ 0
30y1 (ω1 ) − 10y2 (ω1 ) + 30y1 (ω2 ) − 10y2 (ω2 ) − 2x + ν1 ≥ 0 *******
30y1 (ω1 ) − 10y2 (ω1 ) + 30y1 (ω2 ) − 10y2 (ω2 ) − 2x + ν2 ≥ 0
x ∈ {0, 1}, y1 (ω1 ), y2 (ω1 ), y1 (ω2 ), y2 (ω2 ) ∈ {0, 1}

26 / 31
Introduction and Motivation
Closely Related Work
Methodology
Example Illustration
Summary

k = 1: Feasible regions for scenarios 1 (ω1 ) and 2 (ω2 ) at x = 0 and


x =1

27 / 31
Introduction and Motivation
Closely Related Work
Methodology
Example Illustration
Summary

Cut at k=1

28 / 31
Introduction and Motivation
Closely Related Work
Methodology
Example Illustration
Summary
k = 2: Updated feasible regions for scenarios 1 (ω1 ) and 2 (ω2 ) at x = 0
and x = 1

Cut at k=1
Cut at k=2

29 / 31
Introduction and Motivation
Closely Related Work
Methodology
Example Illustration
Summary

Summary
 Mean-risk SMIPs are difficult to solve and easily becoming intractable for
standard solvers.
 In this talk, we present the Fenchel disjunctive decomposition (FDD)
method for SMIP with absolute semideviation which combines subgradient
optimization, Fenchel decomposition and disjunctive programming
 The FDD cuts take advantage of the fixed recourse property and are
generated based only on a subset of scenarios via the common-cut
-coefficient approach.
 Fenchel based cuts are generally expensive to compute are better suited for
problems with special structure such as fixed recourse involving knapsack
constraints and first- and second-stage binary decision variables
 Practical applications and preliminary computational results motivated the
derivation of the FDD method – implementation is underway!

30 / 31
Introduction and Motivation
Closely Related Work
Methodology
Example Illustration
Summary

Questions and Comments?

Thank You!

31 / 31

You might also like