operational research & managementoperations scheduling economic lot scheduling 1.summary machine...
TRANSCRIPT
Operational Research & Management Operations Scheduling
Economic Lot Scheduling
1. Summary Machine Scheduling
2. ELSP (one item, multiple items)
3. Arbitrary Schedules
4. More general ELSP
Operational Research & Management Operations Scheduling
Topic 1
Summary Machine Scheduling Problems
Operational Research & Management Operations Scheduling 3
Machine Scheduling
Single machine
Parallel machine
jjCw||1
max1|| C
1| , |j jr prmp C
1| |j jr Cmax1| |prec h
max1| , |jr prmp L
max1| |jr L
1|| jU1|| j jw U1|| jT1|| j jw T
max||Pm C
max| |P prec Cmax| |Pm prec C
|| jPm C3-partition
max1|| L1|| jC
max2 ||P C
partition
Operational Research & Management Operations Scheduling 4
Why easy?
Show polynomial algorithm gives optimal solution
Often by contradiction
– See week 1
– See week 1
By induction
– E.g.
jjCw||1
max1| |prec h
1|| jU
Operational Research & Management Operations Scheduling 5
Why hard?
Operational Research & Management Operations Scheduling 6
Why hard?
Operational Research & Management Operations Scheduling 7
Why hard?
Operational Research & Management Operations Scheduling 8
Why hard?
Some problem are known to be hard problems.
– Satisfiability problem
– Partition problem
– 3-Partition problem
– Hamiltonian Circuit problem
– Clique problem
Show that one of these problems is a special case of your problem
– Often difficult proof
Operational Research & Management Operations Scheduling 9
Partition problem
Given positive integer a1, …, at and b with ½ j aj = b
do there exist two disjoint subsets S1 and S2 such that
jS aj = b ?
Iterative solution:
– Start with set {a1} and calculate value for all subsets
– Continue with set {a1, a2} and calculate value for all subsets; Etc.
– Last step: is b among the calculated values?
Example: a = (7, 8, 2, 4, 1) and b = 11
Operational Research & Management Operations Scheduling 10
Some scheduling problems
Knapsack problem
Translation: n = t; pj = aj; wj = aj; d = b; z = b;
Question: exists schedule such that j wjUj z ?
Two-machine problem
Translation: n = t; pj = aj; z = b;
Question: exists a schedule such that Cmax z ?
1| |j j jd d w U
max2 ||P C
Operational Research & Management Operations Scheduling 11
3-Partition problem
Given positive integer a1, …, a3t and b with ¼b < aj < ½b
and j aj = t b do there exist pairwise disjoint three element subsets Si such that
jS aj = b ?
Operational Research & Management Operations Scheduling
Topic 2a
Economic Lot Scheduling Problem:
One machine, one item
Operational Research & Management Operations Scheduling 13
Lot Sizing
Domain:
– large number of identical jobs
– setup time / cost significant
– setup may be sequence dependent
Terminology
– jobs = items
– sequence of identical jobs = run
Applications
– Continuous manufacturing: chemical, paper, pharmaceutical, etc.
– Service industry: retail procurement
Operational Research & Management Operations Scheduling 14
Objective
Minimize total cost
– setup cost
– inventory holding cost
Trade-off
Cyclic schedules but acyclic sometimes better
Operational Research & Management Operations Scheduling 15
Scheduling Decisions
Determine the length of runs
– gives lot sizes
Determine the order of the runs
– sequence to minimize setup cost
Economic Lot Scheduling Problem (ELSP)
Operational Research & Management Operations Scheduling 16
Overview
One type of item / one machine
– without setup time
Several types of items / one machine
– rotation schedules
– arbitrary schedules
– with / without sequence dependent setup times / cost
Generalizations to multiple machines
Operational Research & Management Operations Scheduling 17
Minimize Cost
Let x denote the cycle time
Demand over a cycle = Dx
Length of production run needed = Dx / Q = x
Maximum inventory level = (Q - D) Dx / Q = (Q - D) x
= (1 - ) D x Inventory
Time
x
( )Q D Dx
Q
idle time
Operational Research & Management Operations Scheduling 18
Optimizing Cost
book shorter
Solve
Optimal Cycle Time
Optimal Lot Size
21min
2
c D xh Dx
x Q
2
( )
Qcx
hD Q D
2
( )
cDQgx
h Q D
1min 1
2
ch Dx
x
2
(1 )
cx
h D
2
(1 )
cDgx
h
Operational Research & Management Operations Scheduling 19
With Setup Time
Setup time s
If s x(1-) above optimal
Otherwise cycle length is optimal1
sx
nutilizatioq
g
Operational Research & Management Operations Scheduling
Topic 2b
Economic Lot Scheduling Problem:
Lot Sizing with Multiple Items
With Rotation Schedule
Operational Research & Management Operations Scheduling 21
Multiple Items
Now assume n different items
Demand rate for item j is Dj
Production rate of item j is Qj
Setup independent of the sequence
Rotation schedule: single run of each item
Operational Research & Management Operations Scheduling 22
Scheduling Decision
Cycle length determines the run length for each item
Only need to determine the cycle length x
Expression for total cost / time unit
Operational Research & Management Operations Scheduling 23
Optimal Cycle Length
Average total cost
with
Solve as before
1
nj
jj
ca x
x
1 1
n n
j jj j
x c a
1(1 )
2j j j ja h D
jjjj
ca x
x
Operational Research & Management Operations Scheduling 24
Example
Items 1 2 3 4
_qj 400 400 500 400
_gj 50 50 60 60
_hj 20 20 30 70
_cj 2000 2500 800 0
Operational Research & Management Operations Scheduling 25
Data of example 7.3.1
1 2 3 4c(j) 2000 2500 800 0h(j) 20 20 30 70D(j) 50 50 60 60Q(j) 400 400 500 400
rho(j) 0.125 0.125 0.12 0.15a(j) 437.5 437.5 792 1785t(j) 2.14 2.39 1.01 0.00
5300jjc
3452jja
Operational Research & Management Operations Scheduling 26
Solution
1
1 1
1
1
½ (1 )
7 44 342 10. .50 15. .60 35. .60 5300
8 50 40
3452 5300 1.5353 1.24 months
n n
j j j jj j
x c h g
jj
jj
c
a
Operational Research & Management Operations Scheduling 27
Solution
The total average cost per time unit is
(alternative 1: )
(alternative 2: )
How can we do better than this?
8554$2213162725592155
34522 5300 3452
5300
j
j
aAC
c
2 5300
or 2 3452 1.241.24
AC
Operational Research & Management Operations Scheduling 28
With Setup Times
With sequence independent setup costs and no setup times the sequence within each lot does not matter
Only a lot sizing problem
Even with setup times, if they are not job dependent then still only lot sizing
Operational Research & Management Operations Scheduling 29
Job Independent Setup Times
If sum of setup times < idle time then our optimal cycle length remains optimal
Otherwise we take it as small as possible
n
jj
n
jjs
x
1
1
1
Operational Research & Management Operations Scheduling 30
Job Dependent Setup Times
Now there is a sequencing problem
Objective: minimize sum of setup times
Equivalent to the Traveling Salesman Problem (TSP)
Operational Research & Management Operations Scheduling 31
Long setup
If sum of setups > idle time, then the optimal schedule has the property:
– Each machine is either producing or being setup for production
An extremely difficult problem with arbitrary setup times
Operational Research & Management Operations Scheduling
Topic 3
Arbitrary Schedules
Operational Research & Management Operations Scheduling 33
Arbitrary Schedules
Sometimes a rotation schedule does not make sense
(remember problem with no setup cost)
For the example, we might want to allow a cycle 1,4,2,4,3,4 if item 4 has no setup cost
No efficient algorithm exists
Operational Research & Management Operations Scheduling 34
Problem Formulation
Assume sequence-independent setup
Formulate as a nonlinear program
sequences lot sizesmin min COST
s.t.
demand met over the cycle
demand is met between production runs
Operational Research & Management Operations Scheduling 35
Notation
Setup cost and setup times
All possible sequences
Item k produces in l-th position
Setup time sl, run time (production) tl, and idle time ul
. , kjkkjk sscc
nhjjjS h :),...,,( 21
kjl qqq
l
Operational Research & Management Operations Scheduling 36
Inventory Cost
Let x be the cycle time
Let v be the time between production of k
Total inventory cost for k is
k
lk
l
ll
g
tq
g
tqv
2)(2
1 ll
llll t
g
qgqh
Operational Research & Management Operations Scheduling 37
Mathematical Program
( ) ( )2
, , 1 1
1 1min ( )n
2mi ) (
l l
lS Sl l l l l
ll l
S x u
Qh Q D c
x D
)
1
(
( )
, 1,...,
( ) , 1,...,
( )
k
l
jk k
j I
lj j j l
lj L
j j jS
j
Q D x k n
Qs u l
D
u x
S
s
Subject to
Operational Research & Management Operations Scheduling 38
Two Problems
Master problem
– finds the best sequence
Subproblem
– finds the best production times, idle times, and cycle length
Key idea: think of them separately
Operational Research & Management Operations Scheduling 39
Subproblem (lot sizing)
Subject to
2
, , 1 1
1 1min ( ) ( )
2l l
ll l l l l
lx u l l
Qh Q D c
x D
1
( ) , 1,...,
( )
l
lj j j l
lj L
j j j
j
Qs u l
D
s u x
But first: determine a sequence
Operational Research & Management Operations Scheduling 40
Master Problem
Sequencing complicated
Heuristic approach
Frequency Fixing and Sequencing (FFS)
Focus on how often to produce each item
– Computing relative frequencies
– Adjusting relative frequencies
– Sequencing
Operational Research & Management Operations Scheduling 41
Step 1: Computing Relative Frequencies
Let yk denote the number of times item k is produced in a cycle
We will
– simplify the objective function by substituting
– drop the second constraint sequence no longer important
1( )
2k k k k ka h q g
Operational Research & Management Operations Scheduling 42
Rewriting objective function
2
1 1
2
21 1
1 1
ρ
1 1( )
2
1
n nk
k k k k k k kk kk
n nk
k k k kk k
n nk
k kk
k
k k
Qy h Q D c y
x D
a y c yx
yxa c
y x
item : timeskk y
substitute: , ρk ka
substitute: k
2
, , 1 1
1 1min ( ) ( )
2l l
ll l l l l
lx u l l
Qh Q D c
x D
Assumption: for each item production runs of equal length and evenly spaced
Operational Research & Management Operations Scheduling 43
Mathematical Program
,1 1
mink
n nk k k
y xk kk
a x c y
y x
1
1 ρn
k k
k
s y
x
Subject to
Remember: for each item production runs of equal length and evenly spaced
Operational Research & Management Operations Scheduling 44
Solution
Using Lagrange multiplier:
Adjust cycle length for frequencies
Idle times then = 0
No idle times, must satisfy
kk
kk sc
axy
1
1n
kk
k k k
as
c s
Operational Research & Management Operations Scheduling 45
Step 2: Adjusting the Frequencies
Adjust the frequencies such that they are
– integer
– powers of 2
– cost within 6% of optimal cost
– e.g. such that smallest yk =1
New frequencies and run times
' ', and k ky
Operational Research & Management Operations Scheduling 46
Step 3: Sequencing
Variation of LPT
Calculate
Consider the problem with machines in parallel and jobs of length
List pairs in decreasing order
Schedule one at a time considering spacing
Only if for all machines assigned processing time < then
equal lot sizes possible
''1
'max ,...,max nyyy
'maxy
'k
'ky
' '( , )k ky
'max
x
y
Operational Research & Management Operations Scheduling
Topic 4
Lot Sizing on Multiple Machines
Operational Research & Management Operations Scheduling 48
Multiple Machines
So far, all models single machine models
Extensions to multiple machines
– parallel machines
– flow shop
– flexible flow shop
Operational Research & Management Operations Scheduling 49
Parallel Machines
Have m identical machines in parallel
Setup cost only
Item process on only one machine
Assume
– rotation schedule
– equal cycle for all machines
Operational Research & Management Operations Scheduling 50
Decision Variables
Same as previous multi-item problem
Addition: assignment of items to machines
Objective: balance the load
Heuristic: LPT with k
kk q
g
Operational Research & Management Operations Scheduling 51
Different Cycle Lengths
Allow different cycle lengths for machines
Intuition: should be able to reduce cost
Objective: assign items to machines to balance the load
Complication: should not assign items that favor short cycle to the same machine as items that favor long cycle
Operational Research & Management Operations Scheduling 52
Heuristic Balancing
Compute cycle length for each item
Rank in decreasing order
Allocation jobs sequentially to the machines until capacity of each machine is reached
Adjust balance
Operational Research & Management Operations Scheduling 53
Further Generalizations
Sequence dependent setup
Must consider
– preferred cycle time
– machine balance
– setup times
Unsolved
General schedules even harder!
Research needed :-)
Operational Research & Management Operations Scheduling 54
Flow Shop
Machines configured in series
Assume no setup time
Assume production rate of each item is identical for every machine
Can be synchronized
Reduces to single machine problem
m
iikk cc
1
Operational Research & Management Operations Scheduling 55
Variable Production Rates
Production rate for each item not equal for every machine
Difficult problem
Little research
Flexible flow shop: need even more stringent conditions