No.NSC87-2212-E-110-017 86 8 87 7 !" #$%&'()* + , - . / 0 1 2 3 4 5 6 7 8 9 : ; < = > ? @ A B C D E 7 F D G H AIJ K L K M N O 7 3 4 P Q ; < F D G H A R 3 4 P F D G H A S T U V D E W X P F D G H A 7 Y Z [ \] H ^ _ % ` a b J c d e f g h i jRk l S T M ; m n o p q r 7 G ? s t = > uIv w x 7 y z { | } > m ~I T S T k l W X P F D G H A 7 F D I e [ \ = > u 7 i jR 0 3 4 P F D G H A # w x 7 ; < = > ? @ A S M N 7 L H Rk l ; < = > ? m ; < 3 U b m } 7 I m y z y d 7I < = > ? * E S = > ¡ ¢ 3 £ ¤ ¥ ¦ B 7 § 2I¨ © ¤ ª « ¬ m = > ? URk l m ® ¯ ° 7 ± b ` ² ³ ´ µ z ¶ 7 } ·I¸ m } L 7 ¥ [ \I¹ ¬ m = > ? 7 Rº »I3 4 P Q ; < F D G H A ¼ ½ M 3 4 P F D G H A : ; < = > ? @ A 7 ¾ jR = I 1 2 » A ¿ À l Á m  à 7 Ä Å = > ? J IÆ Ç 0 H x È k l ; < = > ? @ A : 3 4 P 8 9 7 M % É 7 Ê Ë ÌI0 Í s N O µ M Î ÏI S t K > 7 = > uIÀ Ð Ç 3 4 P Q ; < F D G H A Ñ Ò Ó 2 0 B Ô = > ? 7 Õ Ö ( ½ R ABSTRACT
The purpose of this study is to develop a distributed and multilevel technique in the traditional genetic
algorithms in order to establish a faster and more efficient new method, distributed hybrid multilevel genetic algorithm.
Distributed genetic algorithms can slove the problem of traditional sequential genetic algorithms, such as premature convergence, large number of function evaluations, and a difficulty in sentting parameters. By using several concurrent sub-population, distributed genetic algorithms can avoid premature convergence resulting from the single genetic searching environment of sequential genetic algorithms. It is useful to speed up the operation rate of joining timely multilevel optimization with distributed genetic algorithms. Because multilevel optimization can resolve one problem into several smaller subproblems, each subproblem is independent and not interference with one another. Then the subsystem of each level can be connected by sensitivity analysis. So we can solve the entire problem. Because each subproblem contains less variables and constrains, it can achieve the faster converge rate of the entire optimization. Therefore, distributed hybrid genetic multilevel algorithm consists of two advantages of the distributed genetic algorithms and the multilevel optimization methods.
This method can be applied to several design optimization problems of trusses because of the reduction of execution time of distributed computing. In addition, the efficient searching function and results are also identified in this study. In short, the distributed hybrid multilevel genetic algorithms are efficient tools in solving structural optimization problems.
× 7 F D G H A Ø Genetic Al gor ithm sÙ l ( ) J Â Ú 7 Û 4 Ü = > ? S T M Õ Ö 7 U ÝIÞ H x È Y ß ` a b à á d e â ã f ä å æ , i çR D E = > ? @ A T è é @ P UI[ \ ê L ä ë ì e ç = > ?Rå í î F D G H A [ \ ï Iv ð ñ ò 7 = > uIº » B C ó ô @ A ¾ j õ ó ö , i j 7 Q F D G H A (Hybrid Genetic Al gor ithms) ÷ ø ù ú ¤ R Q F D G H A D E F D G H A [ \ 7 Iû ¶ % Ü B Ô = > ? J ± b ü ; xI H ý þ x S R ; < = > ? A ½ M ; ² ³ ´ 7 = > ? I; < 3 U å } 7 I U x ÈR» ÌI H H å × H 7 ó % I » Q F D G H A B C 5 6I 7 3 4 P Q F D G H AI¶ % Ü B Ô 7 = > ? Ó 2 ½ N ORº »I- ¶ Q F D G H A B C ; < = > ? 8 9I` 3 4 P 5 6 8 9I K ½ N 7 = > ? @ AI 3 4 P Q ; < F D G H Av T % Ü B Ô å I¿ - @ A ¶ ¨ % Ü = > ? 7 UIz } D E = > ? @ A K ½ Ò H N O R D E 7 W X P F D G H A Ø Sequential GAsÙ T p nI m m ~ ! " Ç m × # $ b u I » n o p Ø populationÙ % & ' ] { | ] ( ± f ) * ! ! F D G ¹ I+ , Ç } > 7 × # $ b u 7 - !Rû » ) * ê S t ñ ò 7 } > uIû ä } . x ÈRº » å Ç L H I/ 0 0 % c 7 F D ! b IS 1 m = > uØ suboptimalÙI» x T » = > u å # 2I 3 - 7 Í s 4 È [ I T ; < = > ? A ¤ U 1 = > u R 0Q;<FDGHA#3 4P5689I3-7% nop 3 å b 5 n b } · 7 6 n o p ØsubpopulationÙI{k567¤¹F DGHR-.¶[\7=>?B` NDEWXPFDGHA]3 4PFDGHA`34PQFDGHA ¹8}9:R- 12#$% Hb SUN (§;<7(§; p<IT PVM3.3.10 =~Å>34PH ¤§345689R B 9 : FDGHA¶st=>uêÞM Ö7NI û x Y ? @ 7 i j R 3 4 8 9 7 A I F D G H A 7 [ \ x ÈI B ; 7 ( § ; S C 3 4 P F D G H A B ½ N R 0 w D F D ! b è E = > ? A #Ik l E = > ? A T è é @ P UI¸ L 7 1 [ \ uRû % Ü B Ô F I G H I . d · 7 x ÈR; < J 7 A IC æ å ; < = > ? AI S Ê Ë d · 7 x ÈR¨ , k l ; < = > ? A m = > ? ) ; < 3 U å q m } e 5 6 7 = > ? I m = > ? H 5 6 7 J ± b ` ² ³ ´ z ¶ 7 K · ;R- * #I0 ¥ = > ? B I¹ ¡ ¢ 3 £ x I 1 7 = > ¡ ¢ L bId M S T M N N O - * 7 = > ? B ¶ P * 7 Q RIµ k l S ¹ ² $ b 7 I0 T M U ± - * 7 J ± b V (µ ª 7 = > ? B ) IU ¹ ² $ b 7 W X (Violation)) I Y ¹ Ç ¬ m = > ? ) 7 I¸ ; < = > ? ½ M } Ö 7 [ \ ZR[ C T 7 J \ 7 3 4 P Q ; < F D G H A z } l D E 7 W X P F D G H AId : [ \ u ] [ \ I K å M N O R
r ^ - . 7 3 4 P H I k PVM = ~ Å Ô > 7 R z } l _ ` 7 ø ¥ I PVM a 7 M j b þR¸ æ c Id M ) P 7 d e } å @ fI µ d M g PVM H % k 7 h B I A ø 7 ô ô ` i ( § ; q r ø , 7 q ô j k R ;<=>?l-.#M3Uål <R 3 U å m <I] m < T I z n K L 1 [ \ R BC; <BÔ`89I m % Ü B Ô F 3 U o ; } 7 B ÔIv S p x H d p 7 B ÔIq » Ê Ë o ; ð [ \ u 7 x ÈR B C r @ AI S T + , m s  M N O Ò 2 7 @ A R a / . t
1. Shyue-Jian Wu and Pei-Tse how uIntegrated Discrete and Configuration Optimization of Trusses Using
Genetic Algorithmsv. Computers w
Structures. Vol.56. No.6. pp.979-991. 1995
2. Shyue-Jian Wu and Pei-Tse Chow u Steady-state Genetic Algorithms for Discrete Optimization of Trussesv . Computers w Structures. Vol.56. No.4. pp.695-702. 1995
3. Jahn H. Holland. u Adaptation in
natural and artificial systemsv. Ann Arbor: The University of Michigan Press. 1975
4. David E. GoldberguGenetic Algorithms in Search. Optimization. And Machine
Learning.v Addison Wesley Press.
1989
5. S. Rajeev and C.S. Krishnamoorthy u Discrete Optimization of Structure Using Genetic Algorithmsv Journal of the Structural Engineering. Vol.118. No.5. pp.1233-1249. 1992
6. C.-Y Lin and P. Hajela. u Genetic Algorithms in Optimization Problem
with Discrete and Integer Design Variablesv. Engineering Optimization. Vol.19. pp.309-327. 1992
7 D. E. Grierson and W. H. Pak. uOptimial Sizing, Geome-trical and Topological Design Using a Genetic Algorithm v. Structural Optimization. Vol.6. pp.151-159. 1993
8. M. Ohsaki u Genetic Algorithm for
Topology Optimization of Trussesv
Computers w Structures Vol.57. No.2.pp.219-225.1995
9. Shyue-Jian Wu and Pei-Tse Chow u Genetic Algorithms for Solving
Mixed-Discrete Optimization Problemsv. Journal of the Franklin
Institude. Vol.331B. No.4. pp.381-401. 1994
10. J. J. Grefenstette. uPrarllel Adaptive Algorithms for Function Optimizationv.
Technical Report no. CS-81-19. Nashville: Vanderbit University. Computer Science Department. 1981 11. J. P. Cohoon. S. U. Hegde. W. N. Martin.
and D. S. Richards. u Punctuated
Equilibria: A Parallel Genetic
Algorithmv . In Proceedings of the
second international conference on Genetic Algorithms. pp.148-184. 1987
12. Reiko Tanese, u Parallel Genetic
Algorithm for a Hypercube v .In
Proceedings of the Second International Conference on Genetic Algorithms.
pp.177-183. 1987
13 Reiko Tanese. u Distributed Genetic
Algorithmsv .In Proceedings of the
Third International Conference on Genetic Algorithms. pp.434-440. 1989 14. C. B. Pettey. M. R. Lutze. and J. J.
Grefenstette. u A Parallel Genetic
Algorithm v .In Proceedings of the
Second International Conference on Genetic Algorithms. pp.155-161. 1987 15 H. Muhenbein. M. Gorges-Schleuter.
And O. Kramer.uNew Solution to the Mapping Problem of Parallel System:
The Evalution Approach v .Parallel
pp.269-279. 1987
16 H. Muhenbein. u Parallel Genetic
Algorithms Population Genetics and Combinatioral Optimizationv .Lecture Notes in Artificial Intelligence
565.Parallelism. Learning. E
volution. edited by J. D. Becker et al.Springer-Verlag. Berlin. pp.399-405. 1991
17 H. Muhenbein. uAsynchronous Parallel Search by the Parallel Genetic Algorithmv.Proceedings of the Third IEEE Symposium on parallel and Distributed Processing Dallas.Texas. USA. pp.526-533. December 1991 18 James P. Cohoon. Shailesh U. Hegde.
Worthy N. Martin. and Dana S. Richards. uDistributed Genetic Algorithms for
the Floorplan Design Problemv IEEE
Transactions on Computer-Aided Design. Vol. 10. No.4. pp.483-491. April 1991
19 H. -M Voigt and J. Born.uA Structured Distributed Genetic Algorithm for
Function Optimization v Parallel
Computing and Mathmatical Optimization. edited by Manified Gauer et al.Springer-Verlag. Berlin. pp.199-208. 1991
20 T .C. Fogarty and L. Bull.uOptimising Individual Contral Rules and Multiple Communicating Rule-Based Control Systems with Parallel Distributed Genetic Algorithmsv.IEE Proceedings -Control Theory and Applications. Vol.142. No.3. pp.211-215. May 1995
21 Joachim Stender. Parallel Genetic
Algorithms: Thoery and Applications. IOS Press. 1993
22 D. Whitley. uThe Genitor Algorithm
and Selection Pressure: Why rank-based Allocation of reproductive trials is
Bestv .In Proceedings of the third
international Conference on Genetic Algorithms. pp.116-121. 1989
23 J. J. Grefenstette. u Optimization of Control Parameters for Genetic
Algorithms v .IEEE Transactions on
System. Man. and Cybernetics.
SMA-16(1). pp.122-128 1986
24 V. S. Sunderam]G. A. Geist]J. Dongarra
and R. ManchekuThe PVM concurrent
computing system: Evolution ]
experences ] and trends v I Parallel
Computing I Vol.20 pp. 531-545
1994
25 M. Dorigo and V. Maniezzo,uParallel Genetic Algo-rithms: Introduction and
overview of current research v , in
Stender's[21], pp.5-42, 1993 26 xyzIu34PFDGHAl=>? J vI#$%&'() {|:.I1996 27 }~Iu34PQCFDGHAl= >?J vI#$%&' () {|:.I1997 28 J. Sobieszczanski-Sobieski, B. B. James and Riely M. F. uStructural Sizing by Generalized Multilevel Optimizationv AIAA Journal Vol. 25 No. 1 1987 pp. 139-145
