Anda di halaman 1dari 20

An Mixed Integer Approach for

Optimizing Production Planning


Stefan Emet
Department of Mathematics
University of Turku
Finland
WSEAS Puerto de la Cruz 15-17.12.2008
Outline of the talk
Introduction
Some notes on Mathematical Programming
Chromatographic separation the process behind the model
MINLP model for the separation problem
Objective - Maximizing profit under cyclic operation
PDA constraints

Numerical solution approaches
MINLP methods and solvers
Solution principles
Some advantages and disadvantages

Some example problems
Solution results - Some different separation sequences

Summary
Conclusions and some comments on future research issues
WSEAS Puerto de la Cruz 15-17.12.2008
Optimization problems are usually classified as follows;
Variables Functions
continuous:
masses,
volumes, flowes
prices, costs etc.
discrete:
binary {0, 1}
integer {-2,-1,0,1,2}
discrete values
{0.2, 0.4, 0.6}
linear non-linear
non-convex
quasi-convex
pseudo-convex
convex
Classification of optimization problems...
WSEAS Puerto de la Cruz 15-17.12.2008
variables
continuous integer mixed
l
i
n
e
a
r

n
o
n
l
i
n
e
a
r

LP ILP MILP
NLP INLP MINLP
On the classification...
WSEAS Puerto de la Cruz 15-17.12.2008
The separationproblem...
H
2
O

C
1

C
2

C
2
C
1

Column 1
A one-column-system:
}
dt t c ) (
1
}
dt t c ) (
2
2
2
z
c
D
z
c
u
t
q
F
t
c
kj
j
kj kj kj
c
c
=
c
c
+
c
c
+
c
c
Goal: Maximize the profits
during a cycle, i.e.
max 1/T*(incomes-costs)
( )
|
.
|

\
|


=
}

1
1
1
) , (
1
max
i i
T
i
t
t
H
t t y dt z t c y
i
i
t
WSEAS Puerto de la Cruz 15-17.12.2008
A two-column-system with three components:
H
2
O

Column 1

Column 2

waste
H
2
O

C
1
C
2
C
3
C
1
C
2
C
3

C
3

C
2

C
1

Waste

(Note 2*3 PDEs)
In general C
PDEs/Column, i.e.
tot. K*C
x
1i2
x
1i1
x
2i1
x
2i2
y
in
1i
y
in
2i
y
1i1
y
1i2
y
1i3
y
2i1
y
2i2
y
2i3
WSEAS Puerto de la Cruz 15-17.12.2008
( )


|
.
|

\
|

= =

=
}

K
k
T
i
i i
in
ki
C
j
t
t
H kj kij j
t t wy dt z t c y p
i
i
1 1
1
1
1
) , (
1
max
t
Price of products
Cycle length
Raw-material costs
y
kij
and y
ki
in
are binary decision variables while t
i
and are continuous ones.
p
j
and w are price parameters. K = number of columns, T = number of time
intervals, C = number of components to be separated.
MINLP model for the SMB process...


Objective function:
WSEAS Puerto de la Cruz 15-17.12.2008
MINLP model for the SMB process...


PDEs for the SMB process:
2
2
1 1
1
z
c
D
z
c
u
t
c
Fc
t
c
c F F
kj
j
kj
C
l
kl
jl kj
kj
C
l
kl jl j
c
c
=
c
c
+
c
c
+
c
c
|
.
|

\
|
+ +

= =
| | |
data) from estimated (e.g. parameters are and , , ,
, , 1 , , , 1 for
j jl j
D u F
K k C j
| |
= =

=
=

+ =
=
) , ( ) , 0 (
) 0 , 0 (
) , ( ) ( ) ( ) 0 , (
1
1
z c z c
c c
z t c t x c t y t c
kj kj
in
j j
K
l
H lj lk
in
j
in
k kj
t
| |

= e
=
=
=

=
=

otherwise. , 0
, 1 , , if , 1
) (
) ( ) (
) ( ) (
1
1
1
T i t t t
t
t x t x
t y t y
i i
i
T
i
i lik lk
T
i
i
in
ki
in
k

o
o
o
Logical functions:
Boundary and initial
conditions:
WSEAS Puerto de la Cruz 15-17.12.2008
MINLP model for the SMB process...

Integral constraints for the pure and unpure components;
) 1 ( ) , (
1
kij
t
t
H kj kij
y M dt z t c m
i
i
+ s
}

Pure components:
}
=

i
i
t
t
H kj kij
dt z t c m
1
) , (
Equality constraints:
Unpure components:
) 1 ( ) , (
1
1

}
=
=
>

C
j l
l
kil
t
t
H kj kij
y M dt z t c m
i
i
WSEAS Puerto de la Cruz 15-17.12.2008
MINLP-formulation summary...
Linear
constraints
Non-linear
constraints
Boundary value
problem
Objective
WSEAS Puerto de la Cruz 15-17.12.2008
MINLP-methods..
Branch and Bound Methods
Dakin R. J. (1965). Computer Journal, 8, 250-255.
Gupta O. K. and Ravindran A. (1985). Management Science, 31, 1533-1546.
Leyffer S. (2001). Computational Optimization and Applications, 18, 295-309.

Cutting Plane Methods
Westerlund T. and Pettersson F. (1995). An Extended Cutting Plane Method for Solving Convex MINLP
Problems. Computers Chem. Engng. Sup., 19, 131-136.
Westerlund T., Skrifvars H., Harjunkoski I. and Porn R. (1998). An Extended Cutting Plane Method for
Solving a Class of Non-Convex MINLP Problems. Computers Chem. Engng., 22, 357-365.
Westerlund T. and Prn R. (2002). Solving Pseudo-Convex Mixed Integer Optimization Problems by
Cutting Plane Techniques. Optimization and Engineering, 3, 253-280.

Decomposition Methods
Generalized Benders Decomposition
Geoffrion A. M. (1972). Journal of Optimization Theory and Appl., 10, 237-260.
Outer Approximation
Duran M. A. and Grossmann I. E. (1986). Mathematical Programming, 36, 307-339.
Viswanathan J. and Grossmann I. E. (1990). Computers Chem. Engng, 14, 769-782.
Generalized Outer Approximation
Yuan X., Piboulenau L. and Domenech S. (1989). Chem. Eng. Process, 25, 99-116.
Linear Outer Approximation
Fletcher R. and Leyffer S. (1994). Mathematical Programming, 66, 327-349.

WSEAS Puerto de la Cruz 15-17.12.2008
NLP-subproblems:
+ relative fast convergenge
if each node can be solved
fast.
- dependent of the NLPs
MINLP-methods (solvers)...
Branch&Bound
minlpbb, GAMS/SBB
Outer Approximation
DICOPT
ECP
Alpha-ECP
MILP
MILP
NLP
NLP NLP
NLP NLP
NLP
MILP and NLP-subproblems:
+ good approach if the NLPs
can be solved fast, and the
problem is convex.
- non-convexities implies
severe troubles
MILP-subproblems:
+ good approach if the
nonlinear functions are
complex, and e.g. if gradients
are approximated
- might converge slowly if
optimum is an interior point of
feasible domain.
WSEAS Puerto de la Cruz 15-17.12.2008
SMB example problems...
(separation of a fructose/glucose mixture)

Problem characteristics:
Columns 1 2 3

Variables
Continuous 34 63 92
Binary 14 27 71

Constraints
Linear 42 78 114
Non-linear 16 32 48

PDE:s involved 2 4 6
WSEAS Puerto de la Cruz 15-17.12.2008
Feed mixture
Collect separated products
Purity requirements:
90% of product 1
90% of product 2.
Recycle
Recycle
[min]
[g/100ml] (0.03, 0.97)
} c
2
feed
(0.14, 0.86)
recycle
(0.9, 0.1)
} c
1
(0.37, 0.63)
} c
2
43.5 57 116124.8
0
4
8
12
WSEAS Puerto de la Cruz 15-17.12.2008
Water
Mixture
Fructose
Recycle 1
Glucose
1
14,9 m
t=57-124.8 min t=43.5 - 57 min
t=57-116 min
t= 0- 43.5 min
116-124.8 min
t=0-43.5 min
WSEAS Puerto de la Cruz 15-17.12.2008
Workload balancing problem...
Decision variables:
y
ikm
=1, if component i is in machine k feeder m.
z
ikm
= # of comp. i that is assembled from machine k and feeder m.

Feeders:
WSEAS Puerto de la Cruz 15-17.12.2008
Optimize the profits during a period :
Objective...
where is the assembly time of the slowest
machine:
K k z t t s
M
m
I
i
ikm ik
,..., 1 , . .
1 1
= >

= =
t

=
t
K
k
k k
Y c
1
max
WSEAS Puerto de la Cruz 15-17.12.2008
constraints...
(slot capacity)
km
M
m
ikm ik
S y s s

=1
i
K
k
M
m
ikm
d z =

= = 1 1
(component to place)
(all components set)
0 s
ikm i ikm
y d z
WSEAS Puerto de la Cruz 15-17.12.2008
PCB example problems...

Problem characteristics:
Machines 3 3 3 3 6 6 6 6

Components 10 20 40 100 100 140 160 180
Tot. # comp. 404 808 1616 4040 4040 5656 6464 7272

Variables
Binary 90 180 360 900 1800 2520 2880 3240
Integer 90 180 360 900 1800 2520 2880 3240

Constraints
Linear 172 332 652 1612 3424 4784 5464 6144

cpu [sec] 0.11 0.03 3.33 2.72 5.47 6.44 11.47 121.7
WSEAS Puerto de la Cruz 15-17.12.2008
Summary...
Though the results are encouraging there are issues to be tackled and/or
improved in a future research (in order to enable the solving of larger problems
in a finite time);
- refinement of the models
- further development of the numerical methods
Some references
Emet S. and Westerlund T. (2007). Solving a dynamic separation problem using MINLP
techniques. Applied Numerical Matematics.
Emet S. (2004). A Comparative Study of Solving Some Nonconvex MINLP Problems, Ph.D. Thesis,
bo Akademi University.
Westerlund T. and Prn R. (2002). Solving Pseudo-Convex Mixed Integer Optimization Problems
by Cutting Plane Techniques. Optimization and Engineering, 3, 253-280.
WSEAS Puerto de la Cruz 15-17.12.2008

Anda mungkin juga menyukai