Parallel Simulated Annealing with Adaptive Temperature Determined by Genetic Algorithm - PowerPoint PPT Presentation

1 / 42
About This Presentation
Title:

Parallel Simulated Annealing with Adaptive Temperature Determined by Genetic Algorithm

Description:

– PowerPoint PPT presentation

Number of Views:111
Avg rating:3.0/5.0
Slides: 43
Provided by: Takes9
Category:

less

Transcript and Presenter's Notes

Title: Parallel Simulated Annealing with Adaptive Temperature Determined by Genetic Algorithm


1
???????????????????????????????????????????
???????? ????? ?????? ??? ??, ?? ??, ?? ??
2
????
?????
LSI????
  • ??????????,??????????????????????.

?????
???????? ?????????? ??????????
??????? ????????????
??????????
???????????
  • ?????????(GA)
  • ??????????????(SA)

???????????
3
Simulated Annealing (SA)
  • ??????????????????
  • ???????????????????
  • ??????
  • ???????? SA??????T ???
  • SA???

1. ??????? 2. ?????? 3. ??????????????? 4.
???? 5. ??????????(?????)
????????????????????
?????????????????
4
SA?????????????????
  • ?????????????????????????
  • ??????????????????????

SA????????
Max Temperature
  • ????(????)
  • ????(????)
  • ???????
  • ???????
  • ??????

Temperature
Min Temperature
Time
????????????????????????
Mark 2000
??????(????)?????????????
5
SA?????????????????
  • ?????????????????????????
  • ??????????????????????

SA????????
Max Temperature
  • ????(????)
  • ????(????)
  • ???????
  • ???????
  • ??????

Temperature
Min Temperature
Time
????????????????????????
Mark 2000
??????(????)?????????????
6
??????(??????????)
- eil101 -
TSP Optimum Topt region
eil101 629 1.1 2.5
kroA200 29368 27 53
lin318 42029 20 39
pr439 107217 44 72
rat575 6773 1.7 3.9
d657 48912 14 27
  • ????????????,???????????.???,??????????????.

7
????
  • ?????????????????????SA???
  • ??????????????????

?????????(GA)???,?????????.
?????????????.

GA for Temperature
Parallel SA with Adaptive Temperature determined
by Genetic Algorithm (PSA/AT(GA))
8
PSA/AT(GA)???????
SA
L ??????
SA
L
L
SA
GA operation
SA
Solution
GA operation
Temperature(??)
E N D
GA operation
SA?1?????GA?1??(??)????
0
SA??? GA??????
1
???,GA???????????
2
?SA??????????
3
9
??????????
- eil101 -
???? ???? ????
????SA
  • ????????????
  • ??????????,???????
  • ????????????,?????????????

10
?????
  • ???(Fitness)?,??????????
  • ??????,????L ????,???E ???????Ek ??????????
  • ???,??????,?????
  • ????,???????????????

11
Genetic Algorithm (GA)
  • ?????????????????
  • ????GA????,?????(??)???
  • ???GA Eshelman93 ???

GA?????
Individual
?????????????
Selection
Crossover
????????
Mutation
???????
Population
12
??(Selection)
  • ?????????????????.
  • ????,??????????????.
  • ??????????,?????????????.
  • ?????????SA?????,GA??????????????????.

Important temperature
Low temperature
High temperature
Energy
Energy
Energy
Baseline
Steps
Steps
Steps
Fitness high
Fitness low
Fitness low
13
??(Crossover)
  • BLX-a???????GA???

BLX-a
???(p1,p2)???d ?????ad ?????????????????????
???????????????, ??T ???????X ????.
14
????(Mutation)
  • ????????,??????????,????????????.

????
??µ ??????????sßd
(0ltßlt1)
(???)
d
15
GA??????
16
PSA/AT(GA)???????
SA
SA
SA
GA operation
SA
Solution
GA operation
Temperature(??)
E N D
GA operation
L
SA?1?????GA?1??????
0
L
SA??? GA??????
1
L ??????
???,GA???????????
2
?SA??????????
3
???????????????????
17
????
  • ??????????????(TSP)
  • ?????????????????(JSP)
  • ????
  • PSA/AT(GA)(????)
  • ???????SA(Parallel SAPSA)
  • ????SA(Temperature Parallel SA TPSA)
  • ????
  • ????? ????????()
  • ?????????????????????

18
????
  • ???????SA(PSA)??????????? ??SA??????????
    ?????????????????????????
  • ????SA(TPSA)???????????????,
    ???????????????????????????????????

Steps
19
?????
????? TSP JSP
????? 32 32
???? ???20 1000
1??????????? ????160 32000
???? 50 50
???? ????????50?????????? ????????50??????????
???? ????????????1???????? ????????????1????????
20
?????????????(TSP)
???????PSA/AT(GA)?, ?????????TPSA????,PSA?????????
???
21
??????????????
PSA ???????????????????????????? TPSA????????
??,???????? PSA/AT(GA)??????,??????????
- pr439 -
Error ratio()
5
Num. of annealing steps
(10 )
22
????????(?????)
- eil101 -
Temperature
Num. of annealing steps
Num. of annealing steps
Num. of annealing steps
PSA/AT(GA)??????????TPSA
?????,?????????? PSA ???????????????
23
????????(??????????)
- eil101 -
Temperature
Num. of annealing steps
Num. of annealing steps
Num. of annealing steps
PSA/AT(GA)??????????TPSA
?????,?????????? PSA ???????????????
24
????????(??????????)
- eil101 -
Temperature
Num. of annealing steps
Num. of annealing steps
Num. of annealing steps
PSA/AT(GA)????????,?????????
25
?????????????(JSP)
???????PSA/AT(GA)????????,JSP??????????.
26
? ? ?
????
?????????????????????????? ????????????????(PSA/AT
(GA))
PSA/AT(GA)???
  • ???????SA????GA??????
  • ?????????????????

????
  • TSP,JSP?????????????,?????????
  • PSA/AT(GA)?,
  • PSA???,TPSA??????????????
  • PSA???????????

?????????????????????????,?????????.
27
(No Transcript)
28
????
  • SA?????????????
  • ?????(????????)??????
  • ??????????
  • ????????
  • ???????????
  • (????????????)

29
?????????????????(JSP)
  • n????(Job)?m????(Machine)???
  • ?????????????????(?????)
  • ?????????????????(Makespan)????????????????

30
?????????????????(JSP)
  • n????(Job)?m????(Machine)???
  • ?????????????????(?????)
  • ?????????????????(Makespan)????????????????

31
GA??????(ß?)
- kroA200 -
ß??,0.05?????
32
GA??????(???)
- kroA200 -
????,0.3?????
33
GA??????(?????)
- kroA200 -
??????,1/32?????
34
GA??????(?????)
- kroA200 -
??????,32?????
35
Coding of temperature
  • PSA/AT(GA) uses GA to optimize the cooling
    schedule.
  • Individual Temperature on each processor
  • Design variable The exponent of temperature
    function, X
  • The expression of temperature in PSA/AT(GA) is
    suitable for the exponential cooling schedule in
    SA.

Real Value
Bit Array
Encoding
X
X
Temperature 10 X
Decoding
SA cooling method
36
Cooling schedule (eil101)
- TPSA -
- PSA/AT(GA) -
Temperature
Temperature
Num. of annealing steps
Num. of annealing steps
A line a cooling schedule on one SA.
PSA/AT(GA) Convergence on the important
temperature region TPSA All
processes cant always have good search.
The cooling schedule of PSA/AT(GA) is more proper
than TPSAs.
37
Experimental results (Error Rate) in TSPs
Error Rate ()
TSP Problems
38
SA????
  • ??SA??? D.R.Greening,1990
  • Serial-Like ?????????????????????.
    ????????

- Altered generation ?????SA??????, ????????
?????
  • Asynchronous ????????????
  • ???????,????????
  • ???????????SA???

39
?????????
????????????
????? 10???
2.703
?? 10X
??
PSA/AT(GA)?????????
?????????????
???? 10000 ???? 0.01
??????????? -2,4
Time
40
(No Transcript)
41
(No Transcript)
42
(No Transcript)
Write a Comment
User Comments (0)
About PowerShow.com