WPC~# 2BPZCourier 10cpi#|8 gx6X@87X@HP LaserJet IIIHPLASIII.PRSx  @,\EX@22 ZBP#|8 HP LaserJet IIIHPLASIII.PRSx  @,\EX@HeadingChapter HeadingJ d  ) I. ׃  Right ParRight-Aligned Paragraph Numbers>a݅@  I.   X(# 2.t *BibliogrphyBibliography݅X  (# SubheadingSubheading0\ E A.  "^#"1"1 "1":7#7# 7##"" """" ""    ;;; ;6;pT:pT:p T:p::: :T:99 9999 999191 919:T:T: T::6;6; 6;;#|>#| >#|#g$g$fg$Y$ P$M$P$Y$g${$$${$$$$%%%$ $$+$3$7$3{$+g$A(A(fA(3( )('()(3(A(U(l(({(((((((( ((+(3(7l(3U(+A(,,f,, ,,,,,/,F,],{v,,,,,,,, ,,+v,3],7F,3/,+,//f// ///// 0 070{P0g0{000000 {0g0+P03707 03 0+/33f33 33333334{*4A4U4c4l4o4l4c4 U4A4+*4347333+377f77 77777777{88/8=8F8J8F8=8 /88+8377737+7;;f;t; l;h;l;t;;;;;{;; << <$< <<  <;+;3;7;3;+;<;|<;| <;|;<|<| <|<=|=| =|m=#># >##tg$ g$ fg$ Y$ P$ M$ P$ Y$ g$r {$b $Y $V $Y $b $r $ % % % $ $ $ $ $ $ {$ g$ A( A( fA( 3( )( '( )( 3( A(r U(b l(Y (V (Y (b (r ( ( ( ( ( ( ( ( ( l( U( A( , , f, , , , , , ,r /,b F,Y ],V v,Y ,b ,r , , , , , , , v, ], F, /, , / / f/ / / / / / /r 0b 0Y 70V P0Y g0b {0r 0 0 0 0 0 {0 g0 P0 70 0 0 / 3 3 f3 3 3 3 3 3 3r 3b 3Y 4V *4Y A4b U4r c4 l4 o4 l4 c4 U4 A4 *4 4 3 3 3 7 7 f7 7 7 7 7 7 7r 7b 7Y 7V 8Y 8b /8r =8 F8 J8 F8 =8 /8 8 8 7 7 7 7 ; ; f; t; l; h; l; t; ;r ;b ;Y ;V ;Y ;b 4&P4&Z4&^4&Z4&P4k&>4[&)4P&4L&3P&3[&3k&3&3&3&3&7&7&V7&7&7&7&8&8&(8&38&88&38&(8k&8[&8P&7L&7P&7[&7k&7&7&7&7&;&;&V;&;&;&;&;&;&<& <&<& <&45*)4**4'*3**35*3E*3Z*3q*3*3*3*3*4*)4*>4*P4*Z4*^4q*Z4Z*P4E*7E*7E*V7E*7Z*7q*7*7*7*7*7*8*8*(8*38*88q*38Z*(8E*85*8**7'*7**75*7E*;E*;E*V;E*;Z*|;q*;*;*;*;*;*;*;*<* <*4.P4w.Z4b.^4K.Z44.P4.>4.)4.4.3.3.3.7.7.V7.74.7K.7b.7w.7.7.7.8.8.(8w.38b.88K.384.(8.8.8.7.7.7.7.;.;.V;.;4.|;K.;b.;w.;.;.;.;.;.4b2P4R2Z4<2^4%2Z42P41>41)41413131317171V71727%27<27R27b27l27p28l28b2(8R238<288%2382(81818171717171;1;1V;1;2|;%2;<2;R2;b2;l2;p2;l2;b245)45453535353536363,63<63F64J6)4F6>4<6P4,6Z46^46Z45P457575V757576767,67<67F67J68F68<6(8,6386886385(85858575757575;5;5V;5;5|;6;6;,6;<6;F6;J6;F6;<6<,6 <6<6 <5<5;5;5;5;5;5;5u05u05Vu05d05N05705 05 05/5/5/6/6/,6 0<6 0F670J6N0F6d0<6u0,6060605u05,5,5V,5,5v,5],5F,53,5!,5,5,6,6!,,63,<6F,F6],J6v,F6,<6,,6,6,6,5,5,9,9V,9,9,9,9,9,:,:v,!:],$:F,!:3,:!,:,9,9,9!,93,9F,9],9v,9,9 09 09V 09 09709N09d09u09090909u0:d0:N0!:70$: 0!: 0:/:/9/9/9/9 093939V393949)49>49P49Z49^49Z49P4:>4:)4!:4$:3!:3:3:39393939397979V7979798989(89389889389(8:8:8!:7$:7!:7:7:7979797979;9;9V;9;9;9;9;9<9 <9<9 <9<:;:;!:;$:;!:;:;:;9|;9;9;9;9(9(9V(9(9(9(9(9(:(:(!:($:l(!:Y(:G(:<(99(9<(9G(9Y(9l(9(9(9(9(<6(<6V(<6(,6(6(6(5(5(5(5(5l(5Y(5G(5<(59(6<(6G(,6Y(<6l(F6(J6(F6(<6$<6$<6V$<6$,6$6$6$5$5$5$5$5$5~$5m$5b$5^$6b$6m$,6~$<6$F6$J6$F6$<6$9$9V$9$9$9$9$9$:$:$!:$$:$!:~$:m$:b$9^$9b$9m$9~$9$9$9$9$9$v=$v=V$v=$=$=$=$=$=$=$=$=$=~$=m$=b$=^$=b$=m$=~$v=$k=$h=$k=$v=G(=G(=VG(=Y(v=l(k=(h=(k=(v=(=(=(=(=(=(=(=(=l(=Y(=G(=<(=9(=<(=G(=!,=!,=V!,=3,v=F,k=],h=v,k=,v=,=,=,=,=,=,=v,=],=F,=3,=!,=,=,=,=!,=/=/=/= 0v= 0k=70h=N0k=d0v=d0v=d0v=Bd0v=u0=0=0=0=u0=d0=N0=70= 0= 0=/=/=/=/=/=3=3=V3=3v=3k=4h=)4k=>4v=P4=Z4=^4=Z4=P4=>4=)4=4=3=3=3=3=3=3=3=7=7=:7=7v=7k=7h=8k=8v=(8=38=88=38=(8=8=8=7=7=7="7=7=7=7=7=7=7=7=;=;=V;=;v=;k=;h=;k=;v=<= <=<= <=<=;=;=;=;=;=;=;=|;=;=;=77 7x777 7777 7777 7%7A%7A%7 A%76V7V7 V7_%7_C6_C6 _C6V8655 5AC666 6$%766 6%7x6x6 x6%7C6C6 C6x5x4x4 x4`5~4~4 ~4}444 4~444 4`555 5C6$C6$C6 $C6544 4$`5A`5A`5 A`54VV5VV5 VV5_`5_~4_~4 _~4Vs444 4A~444 4$~4(( (.(.(.( .(%(A'A' A'K'V'V' V'_'_&_& _&V&V%V% V%_%i&i& i&A&K'K' K'$'.(.( .(((( (x.(xK'xK' xK'''' 'K'i&i& i&&xi&xi& xi&&%% %x%}$}$ }$%%% %$%% %%$%$% $%%$$ $$%A%A% A%$V %V % V %_%%% %A%i&i& i&$&"&"& "&"&7#K'7#K' 7#K'#'$.($.( $.(p$($.($.( $.(S%(%.(%.( %.(5&(P&'P&' P&'%K'$K'$K' $K'm%'$'$' $'$K'$i&$i& $i&$&$i&$i& $i&m%&%i&%i& %i&P&&P&%P&% P&%%%%$%$ %$P&%m%%m%% m%%%$ $$ $$  $$$%$%$% $%m%%$%$% $%$%7#%7#% 7#%#%7#i&7#i& 7#i&#&"'"' "'"'7#.(7#.( 7#.(#("%"% "%"%" %" % " %"%>#$>#$ >#$#%6 %6 % 6 %6%6$6$ 6$6%b7$b7$ b7$7%D8$D8$ D8$8%'9$'9$ '9$9%"9%"9% "9%9%"9i&"9i& "9i&9&"9K'"9K' "9K'9'"9.("9.( "9.(w9(8(8( 8(?8.(\7.(\7.( \7.(7(6(6( 6(z6.(z6K'z6K' z6K'6'\7K'\7K' \7K'7'?8K'?8K' ?8K'8'8&8& 8&?8i&\7i&\7i& \7i&7&6&6& 6&z6i&6&6& 6&6&6'6' 6'6'6%6% 6%6%z6%z6% z6%6%\7%\7% \7%7%?8%?8% ?8%8%D87D87 D878X8"98"98 "989:9"99"99 "999:"9:"9: "9:9:"9o;"9o; "9o;w9;8;8; 8;?8o;\7o;\7o; \7o;7;6;6; 6;z6o;z6:z6: z6:6:\7:\7: \7:7:?8:?8: ?8:8:8:8: 8:?89?88?88 ?888:97:97:9 7:9\78\79\79 \797:6:6: 6:z69z68z68 z686:96:96:9 6:96096X86X8 6X86M86767 676X8b77b77 b777X8'97'97 '979X86:6: 6:6:6:6: 6:6:P&:P&: P&:%:$:$: $:m%:%o;%o; %o;5&;S%;S%; S%;$o;$o;$o; $o;p$;#;#; #;7#o;7#:7#: 7#:#:$:$: $:$:$:$: $:$9$8$8 $8$:9$9$9 $9m%:m%:9m%:9 m%:9$8%8%8 %8P&:9%9%9 %9P&:P&X8P&X8 P&X8%7%7%7 %7m%X8$X8$X8 $X8 $7>#7>#7 >#7#X87#87#8 7#8#:97#97#9 7#9#:":": ":":":9":9 ":9"09"X8"X8 "X8"M8":": ":":f7f7 f7777 7,777 7J777 7g7g6g6 g6%7J6J6 J6%7,6,6 ,6%766 6%7C6C6 C6544 4`5~4~4 ~443434 34~4,4,4 ,4`5,5,5 ,5C6C6C6 C6J5g5g5 g5C6C6C6 C686`5`5 `5V5g4g4 g4`5J4J4 J4`5~4~4 ~4Q4n4n4 n4~4~4~4 ~4s477 7%7g.(g.( g.((J.(J.( J.((,.(,.( ,.((.(.( .(f('' 'K',K',K' ,K'''' 'JK'gK'gK' gK'''' ''&& &&gi&gi& gi&&Ji&Ji& Ji&&,i&,i& ,i&&i&i& i&&%% %%$$ $%%% %3$Q$Q$ Q$%,%,% ,%%%% %J%g%g% g%%%% %% % %  %%n$n$ n$%??4UUUUUUUUUUUUU0  %  _ i B   B i  _    ]i 'B  _]'']_ B 'i ]   %]'N O3'Z]u_uZ3 B i  O    Ni B  _##F#%11#%#%%########''%' >z@    #!'J!]e!w!}!w!_e!J!#! B i @   z >i B  _]'*'*'%*'*/+k+++0,r,,,-';-]V-h-n-h-_V-;-- ,B ,i r, 0, + + k+ /+i *B * **u*_o*u***]*'6'6'%6'6 7\777!8c8889',9]G9X9^9X9_G9,99 8B 8i c8 !8 7 7 \7 7i 6B 6 6x6f6_`6f6x66]6'1#1#1#1==#1#}! }! %}! w!Oe!J!#! 3 Z u@ zu>Z 3O N >z@    #!J!Ne!w!}! **F****3/+Zk+u++0,r,u,Z,3-;-V-h-On- --F-,,r,0,++k+/+***N*u*o* u*O*n- n- n- h-V-;-N-111%%11x6x6Fx66663 7Z\7u77!8c8u8Z839,9G9X9O^9 99F988c8!877\7 7666Nx6f6`6 f6Ox6^9 ^9 ^9 X9G9,9N9===11==%%%%%%3ZNu OuZ33ZuO uNZ3O NN O& ####%%  % B i N    O i B  3Zu_u]Z'3O N']_   % B i > z   @ i B  #!J!e!_w!}!w!e!]J!'#!   @ z> ']_ & %%%%%%  % N O3 Z>uz@ u Z 3 #!J!e!Ow! }!w!e!NJ!#!   @ z>   %  >z@    #! J!>e!zw!}!w!@ e! J! #! #! J! e!@ w!}!w!ze!>J! #!   @ z> * * %* */+k+++0,r,,,- ;->V-zh-n-h-@ V- ;- - ,#!,J!r,e!0,w!+}!+w!k+e!/+J!*#!* * * u*@ o*u**z*>* %%& %%%1%=%=11%1%6 6 %6 6 7\777!8c8889 ,9>G9zX9^9X9@ G9 ,9 9 8#!8J!c8e!!8w!7}!7w!\7e! 7J!6#!6 6 x6 f6@ `6f6x6z6>6 }!+}!+%}!+w!0,e!r,J!,#!, - ;- V-@ h-n-h-zV->;- -,,r,0,++k+/+* *>*z*u*o*@ u* * * *#!*J!/+e!k+w!+}!+*r,*r,F*r,*,*,*-/+;-k+V-+h-+n-0,h-r,V-,;-,--,;-,V-r,h-0,n-+-*-*F-*,*,*r,*0,u*+o*+u*k+*/+******/+*k+u*+o*+u*0,*r,n-+n-+n-+h-+V-k+;-/+-*1%1%1%%%%1111%x6r,x6r,Fx6r,6,6,6- 7;-\7V-7h-7n-!8h-c8V-8;-8-9,,9,G9r,X90,^9+9*9*F9*8*8*c8*!8u*7o*7u*\7* 7*6*6*6/+x6k+f6+`6+f60,x6r,^9+^9+^9+X9+G9k+,9/+9*=%=%=%1%11=1=%%%%%%%%1%1%%%%%%11%**%*3*Z/+uk+++0,ur,Z,3,-;-V-Oh- n-h-V-N;--,,r,0,++k+/+**N**u* o*Ou****%%& %#%#11#1#==1% * *% *B *i /+ k+ + + 0, r,i ,B , -;-V-_h-n-h-V-];-'-,,r,0,++k+/+*'*]**u*o*_u*** * 6 6% 6B 6i 7 \7 7 7 !8 c8i 8B 8 9,9G9_X9^9X9G9],9'988c8!877\7 76'6]6x6f6`6_f6x66 611& 1==%=%11=1166%66N6x6f6 `6Of6x66636Z 7u\777!8uc8Z8389,9G9OX9 ^9X9G9N,9988c8!877\7 7666%6 6>6zx6f6`6@ f6 x6 6 6#!6J! 7e!\7w!7}!7w!!8e!c8J!8#!8 9 ,9 G9@ X9^9X9zG9>,9 988c8!877\7 76*6*6%*6*6/+6k+x6+f6+`60,f6r,x6,6,6-6;- 7V-\7h-7n-7h-!8V-c8;-8-8,9,,9r,G90,X9+^9+X9k+G9/+,9*9*8*8*c8u*!8o*7u*7*\7* 7*6%1%1& %1%=1====1111=11%16666& 6666 76\7x67f67`6!8f6c8x6868686v868696,9 7G9\7X97^97X9!8G9c8,9898898,9c8G9!8X97^97X9\7G9 7,9696868x6c8f6!8`67f67x6\76 766====& ====#=#====J======J=J=JJJ==J  4f]] ]f f $fLL Luf f LL Lu$$ $  ]44 4ff fH  4] ]  ] y Q Q  Q )     $ L L  L u      ) )  ) Q LQ LQ  LQ u) $) $)  $) Q y y  y    ]]A]A ]AAA AAA ALALA LAu     y uy uy  uy L LL LuAA A       , } }  } T Aff f  D D  D # # #  #) D          ) )  ) Q Q }Q } Q }) TQ ,Q , Q ,)     v v  v  Q Q  Q )  d d   d Q Q  Q ) d y d y d  y d     d d d  d A A A  A          y Q Q  Q ) y  y   y  , } }  }y Ty y  y   }} }00 0A}A,A, A,f f  f ,,, ,     f     fd d d  d       ? ?  ? f ? A? A?  A? ? ?  ?  ? ?   ? v Q ? Q ?  Q ? )  ? ?   ?           Wmm mWE W11 1 ~~ ~U~~ ~jUB~B~ B~U B000 0gg g0 01010 10WW W0 0~~ ~U~~ ~WUWW W  mm mE  4    W11 1 mm mmm mjEBB B ~~ ~U Bjj j 1~1~ 1~ U 1!! !1m  jBmBm BmE EE Em    4         )     4 44 4B B  B BB Bjj j II Ij4    ] ]Q ]Q  ]Q )       4        1   4W4W4 W49$y $y  $y      4y 4) 4)  4) Q y y  y Y  AA A4) )  ) Q y Dy D y DY ### #D    y  AA AAA Aff f }} }fTTT T} JJJJJJJ==JJ## #D ffDfD fDH### #D j%% %&&1&1 &1%%W%W %W&%% %&~(~( ~('~(~( ~('WV)WV)W V)W.*.*.* .*V)++ ++-- -,++ ++W.*1.*1 .*1V) ~(~( ~('~(~( ~('j~(B~(B ~(B'~(~( ~('V)V) V).*.*B.*B .*BV)++ ++B-B-B -B,++ ++,, ,->/>/ >/f.00 000B0B 0B0>/B>/B >/Bf.f.jf.j f.j>/>/>/ >/f.-- -,j++ ++jV)jV)j V)j.*.*.* .*V)~(1~(1 ~(1' && &%g&B&B &B%&& &%&& &'a(a( a(8)a(^a(^ a(^8)5a( a(  a( 8) a(!a(! a(!8)"a(n#a(n# a(n#8)F$'F$'F$ 'F$&n#&!&! &!'"' '  ' & &^&^ &^'5'' '&8)8) 8)a(8)%8)% 8)%)h'h' h''%** **++ +,q-^q-^ q-^K.5#/ #/  #/ / 0!0! 0!1"0n#0n# 0n#1F$/F$/F$ /F$#/n##/!#/! #/!/"K."K." K."q-!+!+! +!,"q-n#q-n# q-n#K.F$#/ %#/ % #/ %/%0 %0 % 0 %1%H.%H.% H.%q- %,%,% ,%+ %* %* % * %*%8)%8)% 8)%a( %& %& % & %'%*n#*n# *n#*F$+n#+n# +n#,F$*"*" *"*!* *  * * ** **^** **+^+^ +^+q- q-  q- K. 1 1  1 0 0^0^ 0^1511 10#/#/ #//#/^#/^ #/^/5// /#/// //%00 011%1% 1%111 1111 111j1j 1j1I11 1100 00j00 000101 010 >/1>/1 >/1f. -- -n-,, ,-1+1+1 +1e+,W,W ,W->/>/ >/f.Wf.f. f.>/00 0011 1100 00W1W1W 1W1911 11 2? 2?  2? z3Q4? Q4?  Q4? )56? 6?  6? 67? 7?  7? 8a9? a9?  a9? 9:;? ;?  ;? ;  0 i    B f *      f f ^f    + m u T N  ]< ]6 H H ]< r   @ +  + m * 6H]BB& Bl*l6$Bf<B""v"#B)#WS#l#l#W#B###h#}""t#z#""V#}####B#WM#W##B#""" # #B #"""""""#J###b###G# #G#G#VG#t###@##\#)#")#H#H#]#]""#V###X#@##@#@#@#h#"" ""..J....../;/q/w/S//8/.....e/e/ e/8/e/e/Ze///+/mw/G/&/ //]/]/H/H/]/D/q//@}/+w/}/+}/mS/S/lS/l%S/l/W...h....8/k//h///Ww/lS/lM/W/B,/W...h.../;/&/Y//h///W/Bq/WM/W.. ..::& :;;l=;*;;7;d;B:;;* ;;;]u:|;;::p;;n#n#=n#V##""""""I"s#h#### #8#8#MS#M#8" """#\#sz#^#I#z#\###"""""I"^#sb##### #8#$#8M#8#$ # "###b##s#I##n#t#5#"" "" #  #   # )#8MM"MMM M88 8 8+ 8 M MQ 8      ! ]     ^ o ] !  ^ ^    ' c     ^x i W  s 1  - c 8 8K $ 8 8 8 8 $ 8   8 s s si E  ^     9 x   1 1  1 sf 'f '$f ' ' :( {( ( )) >) >)H )) ( {( :( ' ' t'* t'f 'T ' ' :( {( ( )) ) )){ ))B )Z ))$ ( {( :( ' ' '0 'i ' ' '  ' '$t'$t'& $t'6t'>)l(B(<(f()$t'"t'"t'r"t't#t'z#'"'"t'h#O(#(#(#(#)#))#>)S#>))#))#)"("("(#(##)M#))#))#)#(#(#(V#O("t'.'.'f.'.t'.t'8/'k/'/:(/{(/(/))w/>)S/>)/)).(.{(.:(.'.'.'.'.:(.{(.(,/))M/))/)M/))M/)). M/))q/))/)/))/(/{(/:(Y/'&/'/';/'/'/'/'.'.'#/8#/81#/8.......//\////^n/A///..^.^.....5/h/n/t/t/^J/;/)/.s.1...5/8#/8\/M/M/8/ / /8/$z/8V/8/$// /.^;/;/;/h/sz/1n/n/n/J/ /...::"::;M:M:8;8;M::t':t'& :t';t';>)=;(;(;(7;(d;):t'e/3e/3Be/38/z3.e3.e3.z3.3.3.3.3.z3/z3;/3q/3w/3S/3/z3.z3.z3 .z3.3e/3e/3Ze/3/3/3/?4w/l4G/4&/4 /l4//5//5/5/5//5/l4D/l4q/T4/4}/3w/3}/3}/?4S/l4u:3u:3* u:3|;3;3:3:3p;4;e3;e3;/5u:3#3#3F#3t#3G#z3 #e3"e3"z3"3"3"3"3"z3#z3J#3#3#3b#3##z3"z3"z3 "z3"3"l4"l4R"l4"l4#4V#4#l4#*4#4#3#3#4#T4#4\#4)#4"4)#5#5#/5#/5"l4h#4h#4h#4#T4#433* 333/5He36e34333 3 3F 3 3f z3* e3 e3 z3 3 3 3 3 z30 z3i 3 3 3 3B z3 z3 z3  z3 3 3 3Z 3 3 ?4 l4r l4 T4 4 3 3 3 ?4 l4u 4T 4N l4 /5< /56 5 5 /5< l4r l4%% %$ $ $  $2#X"X" X"!    "" "GG Gp  pp pGeGeG eG=p== =e !! ! pp pG="=" ="e  !! !["2#2# 2# $$$ $%e%e% e%=$=2#=2# =2#e $ $ $  $2#!2#!2# !2# $$$ $% % %  %!$M#$M#$ M#$v"%%$%%$% %$%$$$2#$2# $2#%$ $M#2#M#2# M#2#v" $v"["v"[" v"["M#!M#M# M#v" !!!! !! ["!! !["e["e[" e["=!== =e "" "tt t"    !"M#"M#" M#"v"!! !}!%$ %$  %$ $$!$! $!%$["%["%[" %["%:"%#%# %#% $%%%% %%%%% %  % % %% %%%*%* %*%G$p$p $p%$GM#pM#p M#pv"G!p!p !p Gv"v" v"M#%$%$ %$$%w%w %w%$"$" $"%$,),)* ,),)D)D)?C(o)o)[([(o,))) ))o40 40 :40 4N 5N A50 Y5 n5n5Y5e)5W4?455 W4 ?4?4>?45A5eY5Y5A5 50 40 4 o4o4W4W4o4 40 KN KN & KN K33 KN **:*mQQ*oo3?oo o*oo^o3o:*UU:??!!?:*:o?: 3~0 ~0 f~0 9N N 0 }J}`  9~ `J~0 f0 J`f 9 ~$9 9 9   $   `J0  0 0   0   0 0  0 90 ~ 2 v^AvA__zw#/___v#v+ v+ . v+ + F F _+ zz#_zz2 z _F d d F !w!w/!/!v+ /F /F /F w r(r(& r(( (+ (+ D) t)t)\))+ D)+ D)+ JD)+ (F (F (+ r((r((#(((D)t)_t)zt))t)D)+ (+ (+  (+ (r(+ r(+ . r(+ r([(r(#(((D)t)#)z)t)zt)zt)z\)#)((r(_[([(* [([(+ r( (!(/!)/!\)!t) \) D)!)!)!)!(!( (!(!(!( r(+ )!)! )!\) 4*4*^4*4*4*A5*n52+5+5,n5,A5-4 -4 -4-W4,?4,?4+W42+4*4*o42+W4+W4,o4,4-4-4- 4-4,4-4-4-4-A5,)5-)5-")5-Y5,n5,n5+Y52+)5*4*A5*4*4*4*4*4***=*w*/***v*^2+^+v++,/6,_T,w,w,_-w-/ - --v,v,T,6,,/+_+w++2+w*/***v2+v++++/,_6,wT,,,w-w,/--,,,T,6,/,_+++2+*+G*** *v+,, ,-v6v6nv6G7w'8c88w8_8/9988v8v~8~8888/8_8w8wc8_'8/7^666v6~(6~(62 ~(6g(6O(#7g(#7~(6(6 )6P)6})A7})\7h)7 )6P)6P)6P)6 )6(6~(6(6g(7})7})7 })7P)6})7})7Z})7)\7)y7})7P) 8 )'8('8( 8})9~(9~(8h)8h)9( 8 ) 8P)7})7})y7})\7})y7h)7 ) 8 - -" -** - -m -m--Z6Z6& Z6Z988~88B8B6Z6 Y    XX T ` #\  PButHP#JULY 1991`(#DOC.: IEEE P802.11/91    yxbdddy  8oi  yxXdddy  yO  ` #\\  PRu\P#XPage `(#File: 11CHL174#2P'P#U IEEE P802.11 ă ,@Wireless Access Method and Physical Layer Specification    #Xl2P0kXP#yxZdddy  R  DATE: ` ` *June 25, 1991`(# REVISED: July 4, 1991    l yx dddy  Pi  TITLE: ` ` * STRATEGIES FOR CHANNEL ASSIGNMENT  ; yxu dddy  R4  4 l  AUTHOR: ` ` *Chandos A. Rypinski,  ` ` * 5Chief Technical Officer  ` ` *LACE, Inc.`(#Telephone: 707 765 9627  ` ` *921 Transport Way`-(#Facsimile: 707 762 5328  ` ` *Petaluma, California 94954 USA   yxRdddy #Xl2P0kXP# OVERVIEW  A code division spread spectrum channelized signal has the following special characteristics not usually present in a pure time or frequency division channelized plan:  1)` ` *a lesser signaltointerference ratio is needed for separation of samechannel signals than for narrow spectrum bandwidthefficient signals, and(#`  2)` ` *the noise bandwidth of the recovered signal can be proportional to the information carried rather than the occupied spectrum.(#`  Given that a channelized the physical medium is available, it is necessary to use this flexibility in a useful way for which the following are possibilities:  3)` ` *to provide parallel communication paths for each of many simultaneous users, or(#`  4)` ` *to distinguish between overlapping coverage areas served by different radio Accesspoints.(#`  The strategies to be described use the channelization property to permit simultaneous use of Accesspoints with overlapping radio coverage as in 4). Data transfer channels may be positioned on one of the following criteria:  5)` ` *according to a fixed geographic plan allowing all sites to be used simultaneously, or(#`  6)` ` *dynamically assigned to achieve some form of optimization benefit.(#` RECOMMENDATION  It is recommended that none of the plans described be incorporated in the Standard, excepting possibly the Reference Plan in a model. They are offered as evidence that the airinterface and access protocol elsewhere offered, contains the necessary functions for an indefinitely large and a reasonable size smaller coverage system plans.q(.**:: X` hp x (#%'0*,.8135@8:63) 9 44%Spread Spectrum with low chipping/bit ratio (e.g. 1131) 16 44%Linear, spectrum efficient modulations yx~ dddy statistics are important. Since there is much less fading range relative to average level on wideband relative to narrowband signals, the degree of margin needed is modulation dependent. As an example only, and not a basis for decision, the following Table is the Author's unsupported estimate of the suitability of modulations (assuming DOQPSK for spread spectrum chipping). The above table is more conservative than would be used with voice channels because of a high degree of transmission accuracy required for LAN. CONTINUOUS COVERAGE CHANNEL PLANS  It is commonplace in continuous coverage systems to put down a regular pattern of channel group assignments of 7, 21 or other numbers of channels for hexagon based system plans; and 4, 9 or 16 channels for square based system plans. The assumption of square coverages is introduces no more inaccuracy than must be tolerated in real systems considering uncontrollable obstacle locations. A channel is used only at the geographical locations for which it has been planned. The larger the number, the greater the distance between reused channels under saturated usage. In telephone systems, where channels are trunks, nearly all channels can be used simultaneously and efficiently.P(.**::ԌDDeIn packet data systems where each use of a channel is milliseconds at the longest, and where each user is active less than 1% of the time; there can be serious loss of capacity from inflexible channel assignment algorithms . With a fixed number of available channels, an assignment should only be consumed by actual usage and not by geographic constraints. This philosophy shows the greatest advantage when fewer than 50% of the available coverages areas are used simultaneously, and when the number of channels available for allocation is sufficient for distance isolation at the maximum possible usage density. DDeTwo strategies for dynamic channel allocation are offered which are well suited to short holding time traffic found in packet LAN systems. TIMING DESCRIPTION FOR SEQUENTIAL SETUP CHANNEL ASSIGNMENT  DDeIt is now assumed that one dedicated and separate channel is used to setup use of nine data transfer channels. The setup channel is used sequentially at each coverage area holding for a short time when there is no Request, and longer to process a request. It is desirable that less time be required for setup than for data transmission. DDeThe most important parameters that go into the formulation of these plans are:8@).**::@)#$**::8ԌV@ TABLE II ASSUMED TIME DURATION OF MESSAGE FUNCTIONS ׃ Approximate time required in %sec including propagation delay for g) 44%Invitationtorequest and no response on setup chnl:`!(#100 h) 44%Invitationtorequest and a Request on setup chnl:`!(#200 i) 44%Grant plus 48 octet PDF on data xfr chnl:`!(#600 j) 44%Grant plus 288 octet PDF on data xfr chnl:`@!(#2,800 k) 44%Scan time for 9 Norequest coverages:`!(#900 l) 44%Scan time for 9 Request coverages:`@!(#1,800 m)44%Scan time for 16 Norequest coverages:`@!(#1,600 n) 44%Scan time for 16 Request coverages:`@!(#3,200 yx>dddy a)X number of channels in reuse pattern (e.g. 4, 9, 16...) b)X the peak and average duty cycle usage of each accesspoint c)X the setup time required as a fraction of the average message length d)X the setup time used when there is no request e)X timing restrictions from contiguous system operation Assumed Parameters  For a particular 1 Mb/s medium and messagebased access protocol with a sequentially used setup channel serving nine coverage areas, the parameters for setup and data message transmission functions are shown in Table II above. Accesspoint originated traffic is likely to use 4050% of airtime. If the slotwidth for Invitationtorequest and Request is defined by Stationoriginate, it is the same for Accesspointoriginate. One request opportunity may be lost for each Accesspoint originate message. Therefore, the total transportable traffic is similar regardless of its directional makeup. Rounded numbers are used in the above assumptions which are sufficiently accurate for philosophical description. @).**:: Size and Scale Tradeoffs  DDeThe number of coverages scanned should not be too large, or the delay between access opportunities will be too great. If there are too few (less than 9) coverages in a scan, then more than one scanning group will be required for expanded small systems. In this case, the stepping rate for contiguous scanning groups must be synchronized to reduce interference setup channel interference. DDeA further object of the Dynamic Allocation Plan is to maximize the number of served coverages before synchronization is required. There is a size range for which a single scanning group is suitableperhaps 16 to 25 coverages. REFERENCE SYSTEM PLAN  DDeA reference system plan, using synchronized setup channels could be conceived on a square plan of 3 x 3 coverages with a common setup channel and nine data channels as shown above in Figure 1 and also in Figure 2 below. Assuming that there are many like type plans laid out congruently, and that the timing of the setup channel is a uniform 200 %sec stepping period at 1 Mb/s signaling rate; it is then possible (considering only cochannel interference)8P(.**::0* @ **::8 for all channels to be used simultaneously for data transfer. When message length and scan time are considered, it may not be possible to load all channels simultaneously. If it takes 1.8 millisec to assign 9 coverages, and the messages each make use of the data channel only 0.6 millisec, then the first six assigned data channels will be vacant when the scan returns to the starting point. For short messages, only 33% of the available data channel time will be used. For 100% time utilization, the length of the data transfer use in one scan would have to match the setup time used on the setup channel in one scan. ,With a mixture of message lengths occupying the data channels, there will be unused capacity when the average message length is less than the per channel scan time.Ɣ Layout of the Reference Plan  As shown in Figure 2, the italic numbers within the 25 cells represent the allocation of regular 9channel patterns with one channel per coverage. Any given channel number is reused at spacings of two interposed coverages in four directions. This characteristic is independent of traffic, and it is optimum for 100% use of all channels in all coverages simultaneously. This representation is intended to provide a reference against which other possible plans may be compared. It would be possible to have similar regular plans for 4 and 16 channels as well. With more than nine access points, additional like patterns would be placed congruently as shown in Figure 1 above. @).**:: The possibility of fractional groups is present. A!H H  ??kkPDA15L.PLTh<)ؽ R #Xl2P0kXP#ю  System Example: 25 Coverage, 9 Data Channel, 50% Simultaneously Active Coverages $(#(#(#(#KA'#$ DYNAMIC CHANNEL ALLOCATION PLAN DDeRather than trying to increase utilization, the proposed plans assume that 100% use of all coverage sites simultaneously is not a requirement, and that 25 to 50% of the coverages simultaneously active is sufficient enabling advantages in other performance areas. Suppose now that the same number of channels is assigned consecutively as needed. The assignment process might proceed serially from left to right and ascending order of rows from the bottom left. A better sequence algorithm would recognize and follow successive squares of nine coverages so that pileup of demand in more than 9 consecutive assignments would not cause the separation requirement to be violated. DDeA random distribution of 13 active channels (outof25) assigned in one scan sequence is shown for 9channelsL).**::0*N="**:: '#AL available. The larger block numbers are the assigned channel. The closest reuse distance in this illustration is three intervening coverages (compared with two previously) which would result in lesser cochannel interference. This increased separation occurs because the plan takes advantage of 50% simultaneous coverage utilization by not consuming a channel assignment except at active points. Reducing the Number of Channels Required Reducing the number of channels available makes a minimum channel spacing possibility below design criteria. It would be possible to increase the number of channels in the scanned group holding the percentage simultaneously in use constant. Effect of Message Length on Simultaneously Used Coverages  There is an important difference in the scan algorithm for this plan that is not the case for the reference plan. Without a synchronization requirement, the holding time on one coverage is only 100 %sec when there is no request. The scan time for 13 requests and 12 norequests is 3.8 millisec. An average message length of about 300 %seconds would sustain the 13/25 utilization. Longer messages would increase it. If only one site had a request in one scan, the scan time would 2.6 millisec Also, Accesspointoriginate messages would use less than 100 %sec for setup. p&.**:: CONCLUSIONS  DDeWithin the context of the messagebased access protocol running at 1 Mb/s data rate with at least 10 independent channels, there are circumstances where dynamic channel assignment is advantageous. DDeNine channels assigned in regular pattern is a good plan when synchronization of the stepping of the setup channel is available for systems of more than nine coverages. With a small number of additional coverages, in the range 1025 cells, the need for synchronization can be avoided with dynamic channel assignment. Application of Dynamic Channel Allocation  DDeThis method of operation spreads the capacity of a small number of channels over a larger number of radio coverages providing capacity on demand. The capacity has no tie to a location. There should be radio isolation at the edges from contiguous liketype systems, otherwise additional coordination considerations will exist. DDeGood proportions would be in the range of 20 to 50% of coverages simultaneously usable for 10 to 25 coverages in one group. The number of data channels can be less than the number of simultaneous coverages when there are at least 9 data channels at 50% and fewer at lesser usage factors. 8".**::`'yuP**::8 Independence of Station Logic  Any or all of these algorithms are executed in the infrastructure logic. There is no dependence of the airinterface or the station logic on which of these plans is used. The only requirement is that the Station have sufficient channelization, and that the identity of the setup channel is standardized. The implementation of these logics requires that control messages contain channel designation commands and that the default or initial channel be preagreed. Complete Independent Operation of Coverage Areas  It is possible to have setup and data transfer an independent and parallel process on each channel. This possibility is not considered further because it may require the station to scan channels and make a decision on which channel to initiate operation. It is believed that this is unsuitable logic for moving stations, and that it raises the complexity of the station function unnecessarily. There is also the possibility that it will create a centrally managed handoff function which will create other delay factors. RECOMMENDATION  It is recommended that none of these plans be incorporated in the Standard, excepting possibly the Reference Plan in a model. They are offered as evidence that the airinterface and access protocol elsewhere offered, contains the necessary functions for an indefinitely large and a reasonable size smaller coverage system plans.