 |
Оптимизирующий эксперимент
|
|
|
|
Проведем оптимизирующий эксперимент, для того чтобы система нашла автоматически оптимальное решение и сравним его с уже полученным результатом [1].
TAU EQU 10
ZADEL EQU 3
OHE EQU 20
KRIT VARIABLE (2#N$ABC1+2#N$ABC2)-2#N$OTH-0.1#X$PROST-1#N$OTK
NORM1 VARIABLE (NORMAL(1,40,5))
NORM2 VARIABLE (NORMAL(1,60,10))
ERL3 VARIABLE (GAMMA(1,0,TAU, 3))
RAVNOM VARIABLE (UNIFORM(1,80,120))
GENERATE V$ERL3
ASSIGN 1,0
GATE NU STAN1,BBB
AAA1 SEIZE STAN1
ADVANCE V$NORM1
RELEASE STAN1
UNLINK BUF1,AAA1,1
TRANSFER.04,,DDD
EEE LINK BUF2,FIFO
DDD ASSIGN 1+,1
TEST L P1,2,OTH
TRANSFER,BBB
OTH TERMINATE
BBB GATE NU STAN2,CCC
AAA2 SEIZE STAN2
ADVANCE V$NORM2
RELEASE STAN2
UNLINK BUF1,AAA2,1
TRANSFER.08,EEE,DDD
CCC TEST L CH$BUF1,OHE,OTK
LINK BUF1,FIFO
OTK TERMINATE
GENERATE 50000
savevalue kr,V$KRIT
TERMINATE 1
GENERATE,,,1
FFF MARK
TEST GE CH$BUF2,1
SAVEVALUE PROST+,M1
UNLINK BUF2,ABC1,1
ADVANCE V$RAVNOM
TRANSFER,FFF
ABC1 TERMINATE
GENERATE,,,1
KKK MARK
TEST G CH$BUF2,ZADEL
SAVEVALUE PROST+,M1
UNLINK BUF2,ABC2,1
ADVANCE V$RAVNOM
TRANSFER,KKK
ABC2 TERMINATE
Рис.3. Диалоговое окно генератора оптимизирующего эксперимента
Рис.4. Процедура запуска модели
11/09/10 21:57:57 Model Translation Begun.
11/09/10 21:57:57 Ready.
11/09/10 21:57:57 RSM Procedure registered.
11/09/10 21:57:57 RSM_MOVE Procedure registered.
11/09/10 21:57:57 RSM_FULLPLUSCENTER Procedure registered.
11/09/10 21:57:57 RSM_AUGMENTED Procedure registered.
11/09/10 21:57:57 RSM_CENTERTHEFACTORS Procedure registered.
11/09/10 21:57:57 RSM_GETRESULT Procedure registered.
11/09/10 21:57:57 CHAST Procedure registered.
11/09/10 21:58:02 CONDUCT RSM()
11/09/10 21:58:02 **** Experiment in Progress. ****
11/09/10 21:58:02 Simulation in Progress.
11/09/10 21:58:02 A Simulation in an Experiment has ended. Clock is 50000.000000.
11/09/10 21:58:02 Simulation in Progress.
11/09/10 21:58:02 A Simulation in an Experiment has ended. Clock is 100000.000000.
11/09/10 21:58:02 "Run 1. Yield=1961.381969130512. TAU=10; ZADEL=3;"
11/09/10 21:58:02 Simulation in Progress.
11/09/10 21:58:02 A Simulation in an Experiment has ended. Clock is 50000.000000.
11/09/10 21:58:02 Simulation in Progress.
11/09/10 21:58:02 A Simulation in an Experiment has ended. Clock is 100000.000000.
11/09/10 21:58:02 "Run 2. Yield=1938.857372592189. TAU=10; ZADEL=8;"
11/09/10 21:58:02 Simulation in Progress.
11/09/10 21:58:02 A Simulation in an Experiment has ended. Clock is 50000.000000.
11/09/10 21:58:02 Simulation in Progress.
11/09/10 21:58:02 A Simulation in an Experiment has ended. Clock is 100000.000000.
11/09/10 21:58:02 "Run 3. Yield=-1942.06772808578. TAU=20; ZADEL=3;"
|
|
|
11/09/10 21:58:02 Simulation in Progress.
11/09/10 21:58:02 A Simulation in an Experiment has ended. Clock is 50000.000000.
11/09/10 21:58:02 Simulation in Progress.
11/09/10 21:58:02 A Simulation in an Experiment has ended. Clock is 100000.000000.
11/09/10 21:58:02 "Run 4. Yield=-1513.918579316969. TAU=20; ZADEL=8;"
11/09/10 21:58:02 Simulation in Progress.
11/09/10 21:58:02 A Simulation in an Experiment has ended. Clock is 50000.000000.
11/09/10 21:58:02 Simulation in Progress.
11/09/10 21:58:02 A Simulation in an Experiment has ended. Clock is 100000.000000.
11/09/10 21:58:02 "Run 5. Yield=1908.306844669816. TAU=15; ZADEL=5.5;"
11/09/10 21:58:02 Simulation in Progress.
11/09/10 21:58:02 A Simulation in an Experiment has ended. Clock is 50000.000000.
11/09/10 21:58:02 Simulation in Progress.
11/09/10 21:58:02 A Simulation in an Experiment has ended. Clock is 100000.000000.
11/09/10 21:58:02 "Run 6. Yield=1950.7224334872. TAU=15; ZADEL=5.5;"
11/09/10 21:58:02 Simulation in Progress.
11/09/10 21:58:02 A Simulation in an Experiment has ended. Clock is 50000.000000.
11/09/10 21:58:02 Simulation in Progress.
11/09/10 21:58:02 A Simulation in an Experiment has ended. Clock is 100000.000000.
11/09/10 21:58:02 "Run 7. Yield=1913.700634287524. TAU=15; ZADEL=5.5;"
11/09/10 21:58:02 Using Model:
11/09/10 21:58:02 Y = -1374.78 -552.588 A +3477.12 B
11/09/10 21:58:02 +9.01347 A B
11/09/10 21:58:02 -81.1763 A^2 -324.705 B^2
11/09/10 21:58:02 Predicted optimum yield is 1812.81.
11/09/10 21:58:02 Optimum is in the local Experimental Region.
11/09/10 21:58:02 RSM_FitSurfaceToData() returns 4.
11/09/10 21:58:02 Experiment ended.
Оптимальное значение критерия V отражается в матрице LAB_BESTYIELDS, и оно равно 1812,810. Соответствующие этому результату: TAU равен 12,736, а ZADEL равен 5,531.
В результате проведения эксперимента получили следующую функцию регрессии, связывающую реакцию (критерий оптимизации V) с факторами TAU (фактор А) и ZADEL (фактор В):
Y = - 1374,78 – 552,588 A +3477,12 B + 9,01347 A B – 81,1763 A^2 – 324,705 B^2
Проведем эксперимент для полученных значений:
TAU = 15
ZADEL = 6
OHE = 20
START TIME END TIME BLOCKS FACILITIES STORAGES
0.000 50000.000 41 2 0
NAME VALUE
AAA1 4.000
AAA2 15.000
ABC1 33.000
ABC2 41.000
BBB 14.000
BUF1 10010.000
BUF2 10008.000
CCC 20.000
DDD 10.000
EEE 9.000
ERL3 10006.000
FFF 27.000
KKK 35.000
KR 10013.000
KRIT 10003.000
NORM1 10004.000
NORM2 10005.000
OHE 20.000
OTH 13.000
OTK 22.000
PROST 10011.000
RAVNOM 10007.000
STAN1 10009.000
STAN2 10012.000
TAU 15.000
ZADEL 6.000
LABEL LOC BLOCK TYPE ENTRY COUNT CURRENT COUNT RETRY
1 GENERATE 1094 0 0
2 ASSIGN 1094 0 0
3 GATE 1094 0 0
AAA1 4 SEIZE 747 0 0
5 ADVANCE 747 0 0
6 RELEASE 747 0 0
7 UNLINK 747 0 0
8 TRANSFER 747 0 0
EEE 9 LINK 1086 91 0
DDD 10 ASSIGN 71 0 0
11 TEST 71 0 0
12 TRANSFER 63 0 0
|
|
|
OTH 13 TERMINATE 8 0 0
BBB 14 GATE 508 0 0
AAA2 15 SEIZE 410 0 0
16 ADVANCE 410 0 0
17 RELEASE 410 0 0
18 UNLINK 410 0 0
19 TRANSFER 410 0 0
CCC 20 TEST 192 0 0
21 LINK 192 0 0
OTK 22 TERMINATE 0 0 0
23 GENERATE 1 0 0
24 SAVEVALUE 1 0 0
25 TERMINATE 1 0 0
26 GENERATE 1 0 0
FFF 27 MARK 501 0 0
28 TEST 501 0 0
29 SAVEVALUE 501 0 0
30 UNLINK 501 0 0
31 ADVANCE 501 1 0
32 TRANSFER 500 0 0
ABC1 33 TERMINATE 501 0 0
34 GENERATE 1 0 0
KKK 35 MARK 494 0 0
36 TEST 494 0 0
37 SAVEVALUE 494 0 0
38 UNLINK 494 0 0
39 ADVANCE 494 1 0
40 TRANSFER 493 0 0
ABC2 41 TERMINATE 494 0 0
FACILITY ENTRIES UTIL. AVE. TIME AVAIL. OWNER PEND INTER RETRY DELAY
STAN1 747 0.597 39.945 1 0 0 0 0 0
STAN2 410 0.496 60.502 1 0 0 0 0 0
USER CHAIN SIZE RETRY AVE.CONT ENTRIES MAX AVE.TIME
BUF2 91 0 47.680 1086 91 2195.234
BUF1 0 0 0.060 192 3 15.569
SAVEVALUE RETRY VALUE
PROST 0 983.181
KR 0 1875.682
FEC XN PRI BDT ASSEM CURRENT NEXT PARAMETER VALUE
1098 0 50004.141 1098 0 1
3 0 50004.194 3 31 32
4 0 50024.576 4 39 40
1099 0 100000.000 1099 0 23
Таблица характеристик процесса обслуживания
Максимальное значение критерия эффективности V будем искать, изменяя параметр TAU и величину задела [1].
V = С1 N обр - С2 N отк - С3(Тпр3+ Тпр4) - С4 N отн,
где:
Nобр –количество обработок
Nотк – количество отказов
Тпр3 - время простоя первого станка
Тпр4 - время простоя второго станка
С1, С2, С3,С4 – весовые коэффициенты
Весовые коэффициенты: 
Таблица 1 Таблица характеристик процесса обслуживания
| tau | zadel | abc | otk | oth | prost | V |
| 5 | 3 | 998 | 1347 | 7 | 301,861 | -2202,9305 |
| 5 | 997 | 1363 | 6 | 374,428 | -2288,214 | |
| 8 | 1001 | 1369 | 5 | 509,681 | -2364,8405 | |
| 10 | 3 | 997 | 0 | 6 | 458,7 | 1758,65 |
| 5 | 995 | 0 | 6 | 532,1 | 1717,95 | |
| 8 | 1000 | 0 | 7 | 618,315 | 1683,8425 | |
| 15 | 3 | 1001 | 0 | 5 | 624,631 | 1684,6845 |
| 5 | 994 | 0 | 8 | 940,56 | 1509,72 | |
| 8 | 987 | 0 | 4 | 977,747 | 1481,1265 | |
| 17 | 3 | 969 | 0 | 7 | 3357,478 | 252,261 |
| 5 | 981 | 0 | 2 | 1926,388 | 996,806 | |
| 8 | 982 | 0 | 7 | 1405,702 | 1254,149 | |
| 20 | 3 | 831 | 0 | 4 | 17676,305 | -7180,1525 |
| 5 | 853 | 0 | 3 | 14555,111 | -5574,5555 | |
| 8 | 832 | 0 | 4 | 16842,063 | -6761,0315 | |
| 25 | 3 | 680 | 0 | 6 | 32682,004 | -14987,002 |
| 5 | 681 | 0 | 4 | 31300,383 | -14292,1915 | |
| 8 | 663 | 0 | 2 | 33400,359 | -15376,1795 | |
| 30 | 3 | 565 | 0 | 2 | 40937,576 | -19340,788 |
| 5 | 561 | 0 | 7 | 43501,113 | -20635,5565 | |
| 8 | 582 | 0 | 6 | 38063,421 | -17873,7105 |
Максимальное значение критерия эффективности (V=1712,65) было достигнуто при TAU = 16, а величина задела равна 3.
Графики зависимостей
Рис.5. Зависимость критерия эффективности от величины задела.
Рис.6. Зависимость критерия эффективности от интервала поступления.
Рис.7. Зависимость количества раз, когда станок занят от интервала поступления.
Рис.8. зависимость коэффициента использования станков от интервала поступления.
|
|
|
|
|
|
