chapter8.ps

收集了遗传算法、进化计算、神经网络、模糊系统、人工生命、复杂适应系统等相关领域近期的参考论文和研究报告

	} bind def
/COMMONBITMAP { 
	/r exch def
	/d exch def
	gsave
	translate rotate scale /h exch def /w exch def
	/bitmapsave save def 
	r                    
	/is w d wbytes string def
	/cf currentfile def 
	w h d [w 0 0 h neg 0 h] 
	{cf is readhexstring pop} image
	bitmapsave restore 
	grestore
	} bind def
	/proc1 FMLOCAL
	/proc2 FMLOCAL
	/newproc FMLOCAL
/Fmcc {
    /proc2 exch cvlit def
    /proc1 exch cvlit def
    /newproc proc1 length proc2 length add array def
    newproc 0 proc1 putinterval
    newproc proc1 length proc2 putinterval
    newproc cvx
} bind def
/ngrayt 256 array def
/nredt 256 array def
/nbluet 256 array def
/ngreent 256 array def
	/gryt FMLOCAL
	/blut FMLOCAL
	/grnt FMLOCAL
	/redt FMLOCAL
	/indx FMLOCAL
	/cynu FMLOCAL
	/magu FMLOCAL
	/yelu FMLOCAL
	/k FMLOCAL
	/u FMLOCAL
/colorsetup {
	currentcolortransfer
	/gryt exch def
	/blut exch def
	/grnt exch def
	/redt exch def
	0 1 255 {
		/indx exch def
		/cynu 1 red indx get 255 div sub def
		/magu 1 green indx get 255 div sub def
		/yelu 1 blue indx get 255 div sub def
		/k cynu magu min yelu min def
		/u k currentundercolorremoval exec def
		nredt indx 1 0 cynu u sub max sub redt exec put
		ngreent indx 1 0 magu u sub max sub grnt exec put
		nbluet indx 1 0 yelu u sub max sub blut exec put
		ngrayt indx 1 k currentblackgeneration exec sub gryt exec put
	} for
	{255 mul cvi nredt exch get}
	{255 mul cvi ngreent exch get}
	{255 mul cvi nbluet exch get}
	{255 mul cvi ngrayt exch get}
	setcolortransfer
	{pop 0} setundercolorremoval
	{} setblackgeneration
	} bind def
	/tran FMLOCAL
/fakecolorsetup {
	/tran 256 string def
	0 1 255 {/indx exch def 
		tran indx
		red indx get 77 mul
		green indx get 151 mul
		blue indx get 28 mul
		add add 256 idiv put} for
	currenttransfer
	{255 mul cvi tran exch get 255.0 div}
	exch Fmcc settransfer
} bind def
/BITMAPCOLOR { 
	/d 8 def
	gsave
	translate rotate scale /h exch def /w exch def
	/bitmapsave save def 
	colorsetup
	/is w d wbytes string def
	/cf currentfile def 
	w h d [w 0 0 h neg 0 h] 
	{cf is readhexstring pop} {is} {is} true 3 colorimage 
	bitmapsave restore 
	grestore
	} bind def
/BITMAPCOLORc { 
	/d 8 def
	gsave
	translate rotate scale /h exch def /w exch def
	/lb w d wbytes def 
	sl lb lt {lb ms} if 
	/bitmapsave save def 
	colorsetup
	/is im 0 lb getinterval def 
	ws 0 lb getinterval is copy pop 
	/cf currentfile def 
	w h d [w 0 0 h neg 0 h] 
	{ip} {is} {is} true 3 colorimage
	bitmapsave restore 
	grestore
	} bind def
/BITMAPTRUECOLORc { 
        gsave
        translate rotate scale /h exch def /w exch def
        /bitmapsave save def 
        
        /is w string def
        
        ws 0 w getinterval is copy pop 
        /cf currentfile def 
        w h 8 [w 0 0 h neg 0 h] 
        {ip} {gip} {bip} true 3 colorimage
        bitmapsave restore 
        grestore
        } bind def
/BITMAPTRUECOLOR { 
        gsave
        translate rotate scale /h exch def /w exch def
        /bitmapsave save def 
        /is w string def
        /gis w string def
        /bis w string def
        /cf currentfile def 
        w h 8 [w 0 0 h neg 0 h] 
        { cf is readhexstring pop } 
        { cf gis readhexstring pop } 
        { cf bis readhexstring pop } 
        true 3 colorimage 
        bitmapsave restore 
        grestore
        } bind def
/BITMAPTRUEGRAYc { 
        gsave
        translate rotate scale /h exch def /w exch def
        /bitmapsave save def 
        
        /is w string def
        
        ws 0 w getinterval is copy pop 
        /cf currentfile def 
        w h 8 [w 0 0 h neg 0 h] 
        {ip gip bip w gray} image
        bitmapsave restore 
        grestore
        } bind def
/ww FMLOCAL
/r FMLOCAL
/g FMLOCAL
/b FMLOCAL
/i FMLOCAL
/gray { 
        /ww exch def
        /b exch def
        /g exch def
        /r exch def
        0 1 ww 1 sub { /i exch def r i get .299 mul g i get .587 mul
			b i get .114 mul add add r i 3 -1 roll floor cvi put } for
        r
        } bind def
/BITMAPTRUEGRAY { 
        gsave
        translate rotate scale /h exch def /w exch def
        /bitmapsave save def 
        /is w string def
        /gis w string def
        /bis w string def
        /cf currentfile def 
        w h 8 [w 0 0 h neg 0 h] 
        { cf is readhexstring pop 
          cf gis readhexstring pop 
          cf bis readhexstring pop w gray}  image
        bitmapsave restore 
        grestore
        } bind def
/BITMAPGRAY { 
	8 {fakecolorsetup} COMMONBITMAP
	} bind def
/BITMAPGRAYc { 
	8 {fakecolorsetup} COMMONBITMAPc
	} bind def
/ENDBITMAP {
	} bind def
end 
	/ALDsave FMLOCAL
	/ALDmatrix matrix def ALDmatrix currentmatrix pop
/StartALD {
	/ALDsave save def
	 savematrix
	 ALDmatrix setmatrix
	} bind def
/InALD {
	 restorematrix
	} bind def
/DoneALD {
	 ALDsave restore
	} bind def
%%EndProlog
%%BeginSetup
(3.0) FMVERSION
1 1 612 792 0 1 3 FMDOCUMENT
0 0 /Times-Roman FMFONTDEFINE
1 0 /Times-Bold FMFONTDEFINE
2 0 /Times-Italic FMFONTDEFINE
32 FMFILLS
0 0 FMFILL
1 .1 FMFILL
2 .3 FMFILL
3 .5 FMFILL
4 .7 FMFILL
5 .9 FMFILL
6 .97 FMFILL
7 1 FMFILL
8 <0f1e3c78f0e1c387> FMFILL
9 <0f87c3e1f0783c1e> FMFILL
10 <cccccccccccccccc> FMFILL
11 <ffff0000ffff0000> FMFILL
12 <8142241818244281> FMFILL
13 <03060c183060c081> FMFILL
14 <8040201008040201> FMFILL
16 1 FMFILL
17 .9 FMFILL
18 .7 FMFILL
19 .5 FMFILL
20 .3 FMFILL
21 .1 FMFILL
22 0.03 FMFILL
23 0 FMFILL
24 <f0e1c3870f1e3c78> FMFILL
25 <f0783c1e0f87c3e1> FMFILL
26 <3333333333333333> FMFILL
27 <0000ffff0000ffff> FMFILL
28 <7ebddbe7e7dbbd7e> FMFILL
29 <fcf9f3e7cf9f3f7e> FMFILL
30 <7fbfdfeff7fbfdfe> FMFILL
%%EndSetup
%%Page: "144" 1
%%BeginPaperSize: Letter
%%EndPaperSize
612 792 0 FMBEGINPAGE
108 72 540 81 R
7 X
0 K
V
0 12 Q
0 X
(144) 324 73 T
108 90 540 648 R
7 X
V
0 X
(CHAPTER VIII) 284.86 640 T
(SUMMAR) 235.38 604 T
(Y AND CONCLUSIONS) 288.03 604 T
1 F
(8.0  Summary and Conclusions) 108 562 T
0 F
1.16 (The methodology of knowledge-based AI relies on a central assumption concerning) 126 536 P
(the nature of intelligence:) 108 518 T
0.08 (The [knowledge-based AI] paradigm has traditionally assumed that) 162 488 P
2.61 (the symbolic, algorithmic character of the macrophenomena [of) 162 470 P
0.61 (intelligence] also characterizes the inner workings of the cognitive) 162 452 P
2.52 (processor that generates the macrophenomena. \050Chandrasekaran,) 162 434 P
(Goel and Allemang 1988, p. 32\051) 162 416 T
0.66 (Thus, knowledge-based AI models intelligence only at an abstract level and assumes the) 108 386 P
0.24 (resulting models adequately describe the underlying process. At this abstract level, which) 108 368 P
1.3 (Newell \0501981, 1982\051 calls the knowledge level, there is only knowledge, actions, goals) 108 350 P
1.62 (and a rational application of the knowledge to satisfy the goals with the actions. Each) 108 332 P
-0.03 (intelligent action realized by a knowledge-based AI program is explainable by some piece) 108 314 P
(of knowledge within the agent that associated the action to a speci\336c context.) 108 296 T
0.46 (The most direct way to implement intelligent behavior then is to create a knowledge-) 126 266 P
0.66 (base that embodies the pertinent aspects of the task environment so it can be placed into) 108 248 P
0.19 (the problem solver) 108 230 P
0.19 (. Once all relevant knowledge is explicit within the problem solver) 197.65 230 P
0.19 (, the) 519.16 230 P
1 (task is reduced to deciding what knowledge to apply when. If appropriate knowledge is) 108 212 P
1.82 (unavailable, the mechanism can simply fall back to an ordered search until the stored) 108 194 P
0.38 (knowledge is again applicable. This is the essential philosophy of weak and strong meth-) 108 176 P
-0.03 (ods in knowledge-based AI and has been shown time and again to be suf) 108 158 P
-0.03 (\336cient for model-) 456.12 158 P
(ing intelligent behavior) 108 140 T
(.) 219.27 140 T
FMENDPAGE
%%EndPage: "144" 2
%%Page: "145" 2
612 792 0 FMBEGINPAGE
108 63 540 702 R
7 X
0 K
V
108 711 540 720 R
V
0 12 Q
0 X
(145) 522.01 712 T
108 90 540 702.56 R
7 X
V
0 X
2.23 (The dif) 126 694.56 P
2.23 (\336culties associated with knowledge-based AI\325) 162.99 694.56 P
2.23 (s straightforward approach to) 392.42 694.56 P
1.28 (modeling intelligence lie not in the recapitulation of intelligent behavior but in the dis-) 108 676.56 P
1.84 (tinctness between the task environment and the problem solving environment. For the) 108 658.56 P
1.05 (problem solver to have direct access to knowledge about the task environment, the task) 108 640.56 P
0.66 (environment must be modeled within the problem solver) 108 622.56 P
0.66 (. This requires a transduction of) 384.14 622.56 P
0.46 (information from its natural form in the task environment into an internal format directly) 108 604.56 P
-0.05 (accessible by the problem solver) 108 586.56 P
-0.05 (. Every nuance of the task environment that is germane to) 263.69 586.56 P
0.17 (solving the problem must be reconstructed within the agent so it can be invoked at appro-) 108 568.56 P
0.3 (priate points. This transduction leads to problems of credit assignment, knowledge acqui-) 108 550.56 P
3.36 (sition, scaling, knowledge indexing and representation design which are the classic) 108 532.56 P
-0.19 (problems of knowledge-based AI. The motivating thesis of this work is that because of the) 108 514.56 P
0.94 (problems associated with explicit knowledge, literal knowledge level models may be an) 108 496.56 P
0.69 (inappropriate level at which to describe a general computational intelligence. A straight-) 108 478.56 P
-0.1 (forward way to remove these problems is to remove their source, namely the insistence on) 108 460.56 P
0.83 (explicit task knowledge within the problem solver) 108 442.56 P
0.83 (. The objective of this dissertation has) 352.82 442.56 P
2.83 (been to investigate the necessity of knowledge-based AI\325) 108 424.56 P
2.83 (s commitment to providing) 401.25 424.56 P
2.07 (explicitly represented knowledge within an agent to produce intelligent computational) 108 406.56 P
(activity) 108 388.56 T
(.) 143.2 388.56 T
0.32 (In Chapter 2, I described evolutionary algorithms, and in Chapter 3 I ar) 126 358.56 P
0.32 (gued that they) 471.4 358.56 P
0.02 (were indicative of a new weak method that increased its informedness about the or) 108 340.56 P
0.02 (ganiza-) 504.7 340.56 P
2.82 (tion of the search space dynamically during the search. Evolutionary weak methods) 108 322.56 P
2.24 (manipulate a population of solutions much as weak methods like best-\336rst and beam) 108 304.56 P
0.2 (search. Evolutionary weak methods distinguish themselves by their use of representation-) 108 286.56 P
1.67 (speci\336c operators, their propagation of abstract features and the processing strength of) 108 268.56 P
0.28 (reproduction. These factors allow them to be task independent but still perform more like) 108 250.56 P
0.49 (strong methods than weak methods. Also in Chapter 3, I identi\336ed a weak form of credit) 108 232.56 P
1.05 (assignment in these algorithms, termed it empirical credit assignment after its empirical) 108 214.56 P
1.48 (nature, and generalized Holland\325) 108 196.56 P
1.48 (s \0501975\051 analysis of genetic algorithms to describe the) 268.99 196.56 P
0.13 (propagation of bene\336cial abstract features through the population. The distinguishing fea-) 108 178.56 P
0.53 (ture of empirical credit assignment is its ability to adapt its search in a problem indepen-) 108 160.56 P
(dent manner as additional points of the search space are encountered.) 108 142.56 T
1.61 (In Chapter 3, I used the properties of empirical credit assignment to hypothesize a) 126 112.56 P
0.55 (novel sub-branch of AI techniques. Rather than assuming that all knowledge must reside) 108 94.56 P
FMENDPAGE
%%EndPage: "145" 3
%%Page: "146" 3
612 792 0 FMBEGINPAGE
108 63 540 702 R
7 X
0 K
V
108 711 540 720 R
V
0 12 Q
0 X
(146) 522.01 712 T
108 91.12 540 703.12 R