Transcript Kapitel 3
Chapter 1
Strategic Problems: Location
Location problems
Production site
transportation
PS 1 Central warehouse
transportation
CW 1 Distribution centers DC 1
transportation
customers C 1 PS DC C 2 2 CW 2 2 PS C 3 3 DC 3 PS C 4 4 DC 4 …
Full truck load
…
FTL or tours
…
tours
… QEM - Chapter 1 Kapitel 3 / 2 (c) Prof. Richard F. Hartl
More levels possible (regional warehouses)
Can be delegated to logistics service providers
Decision problems
Number and types of warehouses Location of warehouses Transportation problem (assignment of customers) (c) Prof. Richard F. Hartl QEM - Chapter 1 Kapitel 3 / 3
Median Problem
Simplest location problem
Represent in complete graph. Nodes
i
are customers with weights
b i
Choose one node as location of warehouse Minimize total weighted distance from warehouse
Definition: Median
directed graph (one way streets…):
σ
(
i
) = ∑
d ij b j
→ min.
undirected graph:
σ
out (
i
) = ∑
d ij b j
σ
in (
i
) = ∑
d ji b j
→ min … out median → min … in median (c) Prof. Richard F. Hartl QEM - Chapter 1 Kapitel 3 / 4
Example: fromDomschke und Drexl (Logistik: Standorte, 1990, Kapitel 3.3.1) 1/4 2 2/0 3 2 3/2 3 2 5 4/3 4 5/1 4 2 3 6/2 Distance between locations (c) Prof. Richard F. Hartl
D=
0 12 2 12 0 10 4 8 11 2 6 9 5 9 12 2 3 0 10 3 8 0 4 7 6 7 4 5 0 3 12 5 10 2 6 0 QEM - Chapter 1 weight
b=
4 0 2 3 1 2 Kapitel 3 / 5
Example: Median
i\j 1 2 1 4*0 0*12 3 2*2 4 5 3*10 6*1 6 2*12 σ out (i) 64 2 4*2 0*0 3 4*12 0*10 2*3 2*0 3*3 7*1 2*5 3*8 4*1 2*10 40 96 i/j 1 2 3 1 2 3 4 5 4*0 4*12 4*2 4*10 24 0*2 0*0 0*3 2*12 2*10 2*0 0*3 2*8 0 8 4 4*4 5 4*8 0*2 0*6 2*5 2*9 3*0 5*1 2*2 3*4 0*1 2*6 35 74 4 5 3*4 1*8 3*2 1*6 3*5 1*9 3*0 1*4 15 0 6 6 6 4*11 0*9 2*12 3*7 3*1 2*0 92 6 σ in (i) 2*11 2*9 2*12 2*7 66 98 56 74 6 53 0 80
City 4 is median since 35+74 = 109 minimal OUT e.g: emergency delivery of goods
(c) Prof. Richard F. Hartl
IN e.g.: collection of hazardous waste
QEM - Chapter 1 Kapitel 3 / 6 6 48 0 20
Related Problem: Center
Median
Node with min total weighted distance (i) j d ij b j → min.
Center
Node with min Maximum (weighted) Distance (i) max j d ij b j → min.
QEM - Chapter 1 (c) Prof. Richard F. Hartl Kapitel 3 / 7
Solution
i\j 1 2 1 4*0 0*12 3 2*2 4 5 3*10 6*1 6 2*12 out (i) 30 2 4*2 0*0 3 4*12 0*10 2*3 2*0 3*3 7*1 2*5 3*8 4*1 2*10 10 48 4 4*4 5 4*8 0*2 0*6 2*5 2*9 3*0 5*1 2*2 3*4 0*1 2*6 6 4*11 0*9 2*12 3*7 3*1 2*0 16 32 44
City1 is center since 30+24 = 54 minimal
i/j 1 2 3 1 4*0 2 3 4*12 4*2 0*2 0*0 0*3 2*12 2*10 2*0 4 0*3 2*8 5 4*10 24 0 8 4 5 3*4 1*8 3*2 1*6 3*5 1*9 3*0 1*4 15 0 6 in (i) 2*11 2*9 2*12 2*7 24 48 24 40 6 24 6 6 0 48 6 48 0 20 (c) Prof. Richard F. Hartl QEM - Chapter 1 Kapitel 3 / 8
Uncapacitated (single-stage) Warehouse Location Problem – LP-Formulation
single-stage WLP:
warehouse W 1 W 2 m customer: C 1 C 2 C 3 C 4 n Deliver goods to
n customers
each customer has given demand Exist:
m potential warehouse
locations Warhouse in location
i
causes fixed costs
f i
Transportation costs
i
j
are
c i
j if total demand of j comes from i.
(c) Prof. Richard F. Hartl QEM - Chapter 1 Kapitel 3 / 9
Problem: How many warehouses?
(many/few high/low fixed costs, low/high transportation costs Where?
Goal:
Satisfy all demand minimize total cost (fixed + transportation) transportation to warehouses is ignored (c) Prof. Richard F. Hartl QEM - Chapter 1 Kapitel 3 / 10
Example: from Domschke & Drexl (Logistik: Standorte, 1990, Kapitel 3.3.1) Solution 1: all warehouses i\j 1 2 3 4 5 1 1 2 7 6 6 2 2 9 6 5 4 3 10 0 1 10 6 4 9 7 5 2 3 5 6 3 3 6 7 6 7 6 10 3 2 5 6 6 7 3 10 f i 5 7 5 6 5 Fixed costs = 5+7+5+6+5 = 28 high Transp. costs = 1+2+0+2+3+2+3 = 13 Total costs = 28 + 13 = 41 Solution 2: just warehouses 1 and 3 i\j 1 3 1 1 7 2 2 6 3 10 1 4 9 5 5 6 3 6 7 10 7 3 5 f i 5 5 Fixed costs = 5+5 = 10 Transp. costs = 1+2+1+5+3+7+3 = 22 Total costs = 10 + 22 =
32
(c) Prof. Richard F. Hartl QEM - Chapter 1 Kapitel 3 / 11
when locations are decided:
transportation cost easy (closest location) Problem: 2 m 1 possibilities (exp…)
Formulation as LP (MIP)
y i
… Binary variable for
i = 1, …, m
:
y i
= 1 if location i is chosen for warehouse 0 otherwise
x ij
… real „assignment“ oder transportation variable für
i = 1, …,m
and
j = 1, …, n
:
x ij
= fraction of demand of customer j devivered from location i.
(c) Prof. Richard F. Hartl QEM - Chapter 1 Kapitel 3 / 12
transportation cost + fixed cost Delivery only from locations i that are built Satisfy total demand of customer j y i is binary x ij non negative (c) Prof. Richard F. Hartl
MIP for WLP
Z
(
x
,
y
)
i
1
j n m
1
c ij x ij
i m
1
f i y i
min
x ij
≤
y i
i = 1, …, m j = 1, …,n
i m
1
x ij
1
y i
{ 0 , 1 }
x ij
0 j = 1, …,n i = 1, …, m For all i and j QEM - Chapter 1 Kapitel 3 / 13
Problem:
m*n real Variablen und
m binary
→ for a few 100 potential locations exact solution difficult → Heuristics
Heuristics:
Construction or Start heuristics (find initial feasible solution)
Add
Drop
Improvement heuristics (improve starting or incumbent solution ) (c) Prof. Richard F. Hartl QEM - Chapter 1 Kapitel 3 / 14
ADD for WLP
Notation: I 0 I o vl I 1
i
Z I:={1,…,m} set of all potential locations set of (finally) forbidden locations (y i = 0 fixed) set of preliminary forbidden locations (y i = 0 tentaitively) set of included (built, realized) locations (y i =1 fixed) reduction in transportation cost, if location i is built in addition to current loc.
total cost (objective) (c) Prof. Richard F. Hartl QEM - Chapter 1 Kapitel 3 / 15
Initialzation
: Determine, which location to build if just one location is built: row sum of cost matrix c i := ∑c ij … transportation cost choose location k with minimal cost c k + f k set I 1 = {k}, I o vl = I – {k} und Z = c k + f k … incumbent solution compute savings of transportation cost ω ij = max {c kj – c ij , 0} for all locations i from I o vl row sum ω i and all customers j as well as … choose maximum ω i Example: first location k=5 with Z:= c 5 + f 5 = 39, I 1 = {5}, I o vl = {1,2,3,4} (c) Prof. Richard F. Hartl QEM - Chapter 1 Kapitel 3 / 16
i\j 1 2 3 4 5 1 1 2 7 6 6 2 2 9 6 5 4 3 10 0 1 10 6 4 9 7 5 2 3 5 6 3 3 6 7 6 7 6 10 3 2 7 3 10 5 6 6 f i 5 7 5 6 5 c i 38 37 37 38 34 f i + c i 43 44 42 44 39 i\j 1 2 3 4 1 5 4 2 2 3 6 5 1 4 5 1 4 4 1 6 7 3 1 ω i 11 14 10 2 5 f i 7 5 6 (c) Prof. Richard F. Hartl QEM - Chapter 1 ω ij is saving in transportation cost when delivering to custonmer j, by opening additional location i. → row sum ω i is total saving in transportation cost when opening additional location i. Kapitel 3 / 17
Iteration:
in each iteration fix as built the location from I o vl , with the largest total saving: Fild potential location k from I o vl , where saving in transportation cost minus additional fixed cost ω k – f k is maximum.
I
1
I
1 {
k
}, I o vl = I o vl – {k} and Z = Z – ω k + f k Also, forbid all locations (finally) where saving in transportation cost are smaller than additional fixed costs For all
i
I
0
vl
with ωi ≤ fi :
I
0
I
0 {
i
}
and I
0
vl
I
0
vl
{
i
} : Update the savings in transpotrtation cost for all locations I o vl and all customers j ω ij = max { ω ij ω kj , 0} (c) Prof. Richard F. Hartl QEM - Chapter 1 Kapitel 3 / 18
Stiopping criterion:
Stiop if no more cost saving are possible by additional locations from I o vl Build locationd from set I 1 .
Total cost Z assignment: x ij = 1 iff
c ij
min {
c hj h
I
1 }
Beispiel: Iteration 1
i\j 1 2 3 1 2 3 4 5 4 2 6 5 4 4 4 1 5 1 6 7 3 1 ω i 11 14 10 2 f i 5 7 5 6 Fix k = 2 1 Forbid i = 4 Because of ω 4 < f 4 location 4 is forbidden finally. Location k=2 is built. Now Z = 39 – 7 = 32 and I o vl Update savings ω ij.
= {1,3}, I 1 = {2,5}, I o = {4}. (c) Prof. Richard F. Hartl QEM - Chapter 1 Kapitel 3 / 19
i\j 1 3 1 1 2 2 3 4 5 6 7 3 1 ω i 6 1 f i 5 5
Iteration 2:
Location 3 is forbidden, location k = 1 is finally built. Fix k = 1 Forbid i = 3
Ergebnis:
Final solution I 1 = {1,2,5}, I o = {3,4} and Z = 32 – 1 = 31.
Build locations 1, 2 and 5 Customers {1,2,7} are delivered from location 1, {3,5} from location 2, and {4,6} from location 5. Total cost
Z = 31
.
(c) Prof. Richard F. Hartl QEM - Chapter 1 Kapitel 3 / 20
DROP for WLP
Die Set I o vl is replaced by I 1 vl .
I 1 vl set of preliminarily built locations (y i =1 tentatively) DROP works the other way round compared to ADD, i.e. start with all locations temporarily built; in each iteration remove one location… Initialisation: I 1 vl = I, I 0 = I 1 = { } Iteration In each Iteration delete that location from most. I 1 vl (finally), which reduces total cost If deleting would let total cost increase, fix this location as built QEM - Chapter 1 Kapitel 3 / 21 (c) Prof. Richard F. Hartl
Expand matrix C: Row m+1 (row m+2) contains smallest c h1j (second smallest c h2j ) only consider locations not finally deleted →
i
I
0 Row m+3 (row m+4) contains row number h 1 smallest) cost elemet occurs. If location h 1 (and h 2 ) where smallest (second (from I 1 vl ) is dropped, transportation cost for customer Kunden j increase by c h2j - c h1j Example: Initialisation and Iteration 1: I 1 vl ={1,2,3,4,5} i\j 1 2 3 4 5 6 7 δ i f i 1 1 2 10 9 6 7 3 5 5 build 2 2 9 0 7 3 6 10 1 7 delete 3 4 7 6 6 5 1 10 5 2 3 6 10 3 5 6 0 1 5 6 5 6 4 6 3 7 2 6 1 5 c h1j c h2j h 1 h 2 6 7 8 9 1 2 1 2 2 4 1 5 0 1 2 3 2 3 4 5 3 3 2 3 2 3 5 4 3 5 1 3 (c) Prof. Richard F. Hartl QEM - Chapter 1 Kapitel 3 / 22
For all i from I 1 vl δ i compute increase in transportation cost δ i if I is finally dropped. is sum of differences between smallest and second smallest cost element in rows where i = h 1 contains the smallest element.
2 examples: δ 1 = (c 21 δ 2 = (c 33 – c 11 ) + (c 52 – c 23 ) + (c 35 – c 12 ) + (c 37 – c 25 ) = 1 – c 17 ) = 5 If fixed costs savings f i exceed additional transportation cost δ i , finally drop i. In Iteration 1 location 1 is finally built.
Iteration 2: I 1 vl = {3,4,5}, I 1 = {1}, I 0 = {2} Omit row 2 because finally dropped. Update remaining 4 rows, where changes are only possible where smallest or second smallest element occurred Keep row 1 since I 1 compute δ i there.
= {1}, but 1 is no candidate for dropping. Hence do not (c) Prof. Richard F. Hartl QEM - Chapter 1 Kapitel 3 / 23
i\j 1 3 4 5 c h1j c h2j h 1 h 2 1 6 1 4 1 1 7 6 6 2 4 1 5 2 2 6 5 4 1 6 3 5 3 10 1 10 6 2 3 4 5 4 9 5 2 3 3 6 3 1 5 6 3 6 7 2 3 5 4 6 7 10 3 2 3 5 1 3 7 3 5 6 6 8 1 δ i 1 f i 5 6 5 build forbid Location 3 is finally built, location 4 finally dropped.
(c) Prof. Richard F. Hartl QEM - Chapter 1 Kapitel 3 / 24
Iteration 3: I 1 vl = {5}, I 1 = {1,3}, I 0 = {2,4} i\j 1 3 5 c h1j c h2j h 1 h 2 1 6 1 5 1 1 7 6 2 4 1 5 2 2 6 4 3 10 1 6 1 6 3 5 3 5 5 3 4 9 5 3 3 6 3 1 5 6 3 7 2 7 5 1 6 7 10 2 3 5 1 3 7 3 5 6 7 δ i f i 5 build Location 5 is finally built (c) Prof. Richard F. Hartl QEM - Chapter 1 Kapitel 3 / 25
Result
:
Build locations I 1 = {1,3,5} Deliver customers {1,2,7} from 1, customers {3,5} from 3, and customers {4,6} from 5. Total cost
Z = 30
(slightly better than ADD – can be the other way round) (c) Prof. Richard F. Hartl QEM - Chapter 1 Kapitel 3 / 26
Improvement for WLP
In each iteration you can do: Replace a built location (from I 1 ) by a forbidden location (from I 0 ). Choose
first improvement
of
best improvement
Using rules of DROP-Algorithm delete 1 or more locations, so that cost decrease most (or increase least) and then apply ADD as long as cost savings are possible. Using rules of ADD-Algorithm add 1 or more locations, so that cost decrease most (or increase least) and then apply DROP as long as cost savings are possible. QEM - Chapter 1 Kapitel 3 / 27 (c) Prof. Richard F. Hartl
P-Median
Number of facilities is fixed … p Typically fixed costs are not needed (but can be considered if not uniform)
(c) Prof. Richard F. Hartl QEM - Chapter 1 28
transportation cost + fixed cost Delivery only from locations i that are built
MIP for
p
-Median
Z
(
x
,
y
)
i m n
1 1
j c ij x ij
min
x ij ≤ y i
i = 1, …, m j = 1, …,n Satisfy total demand of customer j y i is binary x ij non negative Exactly p facilities (c) Prof. Richard F. Hartl
i m
1
x ij
1
y i
{ 0 , 1 }
x ij
0
i m
1
y i
p
QEM - Chapter 1 j = 1, …,n i = 1, …, m For all i and j Kapitel 3 / 29