Changeset 1427
- Timestamp:
- 02/03/12 14:43:10 (13 years ago)
- Location:
- applications/doprava/texty/novotny_vyzk_LQ
- Files:
-
- 17 added
- 1 removed
- 11 modified
Legend:
- Unmodified
- Added
- Removed
-
applications/doprava/texty/novotny_vyzk_LQ/04_Bayes/Bayes.tex
r1425 r1427 1 \s ubsection{Bayesovské učení} \label{sec:bayes}1 \section{Bayesovské učení} \label{sec:bayes} 2 2 3 3 % \subsection{Úvod} -
applications/doprava/texty/novotny_vyzk_LQ/06_Bayes_rmm_pouziti/Bayes_rmm_pouziti.tex
r1424 r1427 1 1 \def \cesta {./06_Bayes_rmm_pouziti} 2 2 3 \s ubsection{Použití RMM a Bayesova učení v decentralizovaném řízení dopravy}3 \section{Použití RMM a Bayesova učení v decentralizovaném řízení dopravy} 4 4 5 5 V následujícím textu se budeme zabývat konkrétním využitím … … 41 41 \end{figure} 42 42 43 \subsection{Zhodnocení}44 V článku \cite{4_rmm_formalization} nenjsou podrobně popsány akce agentů ani45 způsob, jak hodnotit jejich užitečnost. Proto je tato metoda jen obtížně reprodukovatelná,46 modifikovatelná či dále rozvinutelná. V naší situaci popsanné v dalších kapitolách také není nutné47 modelovat chování agentů, neboť je možné ho vykomunikovat pomocí posílaných zpráv. V případě48 reálného nasazení by však bylo možné vylepšení zapojení RMM pro odhad chování agenta49 pokud by nastal výpadek spojení nebo podobná situace.50 43 51 44 -
applications/doprava/texty/novotny_vyzk_LQ/LQ_rizeni.tex
r1424 r1427 34 34 35 35 36 \subsection{Použití LQ řízení ve strategii TUC}37 LQ řízení bylo použito v \cite{6_tuc_lq} k nalezení optimální délky zelených v systému38 13-ti signálních skupin. Proměnné $x_i(t)$ zde představují obsazenost ramene $i$39 spojující křižovatky $M$ a $N$. Účelem strategie je nalezení optimální délky zelených40 $g$, $g_{N,i}$ značí délku zelené na signální skupiny křižovatky $N$ zprůjezdňující41 směr do ramene $i$. Předpokládaný vztah pro přechod systému z času $t$ do času $t+1$ je42 \begin{equation}\label{eq_tuc_1}43 x_i(t+1) = x_i(t) + T [ q_i(t) + s_i(t) + d_i(t) + u_i(t) ] \;,44 \end{equation} kde proměnné značí:45 46 \begin{itemize}47 \item $T$ - časový krok48 \item $q_i(t)$ - přírůstek vozidel z křižovatek49 \item $u_i(t)$ - úbytek vozidel do ostatních křižovatek50 \item $s_i(t)$ - přírůstek vozidel z okolí sítě51 \item $d_i(t)$ - úbytek vozidel mimo síť52 \end{itemize}53 54 Přírůstek vozidel z křižovatek je dán vztahem55 \begin{equation}56 q_i(t) = \sum_{k\in I_m} t_{k,i} u_k(t) \;,57 \end{equation}58 je to tedy součet úbytků vozidel z ramen ústících do křižovatky $N$ vynásobených59 koeficinety $t_{k,i}$, což jsou odbočovací poměry z ramene $k$ do ramene $i$.60 V podovném tvaru se předpokládá $s_i(t)$61 \begin{equation}62 s_i(t) = t_{i,0} q_i(t) \;,63 \end{equation}64 kde $t_{i,0}$ je odbočovací koeficient ramene $i$ mimo sledovanou síť.65 Při délce cyklu $C$, saturovaném toku $S_i$ a délce zelených $g_{N,i}(t)$ ramene $i$ platí66 \begin{equation} \label{eq:tuc_u}67 u_i(t) = \frac{S_i \sum g_{N,i}(t)}{C}68 \end{equation}.69 Rovnice \ref{eq_tuc_1} tedy přechází do tvaru70 \begin{equation}\label{eq_tuc_2}71 x_i(t+1) = x_i(t) + T \left[72 (1-t_{i,0}) \sum_{k\in I_m} t_{k,i} \frac{S_k \sum g_{M,k}(t)}{C}73 - \frac{S_i \sum g_{N,i}(t)}{C}74 + d_i(t) \right]75 \end{equation}. Uvažujeme-li nominální hodnoty $d^n$ a $g^n$ vedoucí76 vždy na stav $x^n$, platí podle rovnice \ref{eq_tuc_2}77 \begin{equation}\label{eq_tuc_nom}78 0 = T \left[79 (1-t_{i,0}) \sum_{k\in I_m} t_{k,i} \frac{S_k \sum g_{M,k}^n}{C}80 - \frac{S_i \sum g_{N,i}^n}{C}81 + d_i^n \right]82 \end{equation}. Označíme-li83 \begin{equation}\label{eq_delta_g}84 \Delta g(t) = g(t) - g^n85 \end{equation}, můžeme psát rovnici \ref{eq_tuc_2} jako86 \begin{equation}\label{eq_tuc_3}87 x_i(t+1) = x_i(t) + T \left[88 (1-t_{i,0}) \sum_{k\in I_m} t_{k,i} \frac{S_k \sum \Delta g_{M,k}(t)}{C}89 - \frac{S_i \sum \Delta g_{N,i}(t)}{C}90 \right]91 \end{equation}, což dovoluje tuto rovnici zapsat pomocí matic v požadovaném tvaru92 \begin{equation}\label{eq_tuc_4}93 x(t+1) = A x(t) + B \Delta g(t)94 \end{equation}, kde $A$ je jednotková matice.95 96 \subsubsection{Kvadratické kritérium}97 Účelem lagoritmu je minimalizovat obsazenost ramen, tedy vektor $x(t)$98 a penalizovat změnu délky trvání zelené oproti nominálním hodnotám.99 Kvadratické kritérium optimálního řízení \ref{eq_quadratic_criterion} jetedy v \cite{6_tuc_lq}100 definováno vztahem101 \begin{equation}\label{eq_tuc_crit}102 J = \sum_{t=0}^{\infty} x(t)^T Q x(t) + \Delta g(t)^T R \Delta g(t)103 \end{equation}. Diagonální matice $Q$ je zde zodpovědná za vyvažování104 počtu vozidel jednotlivých úseků. V \cite{6_tuc_lq} je každý diagonální105 prvek $Q_{i,i}$ matice $Q$ položen převrácené hodnotě maximálního106 povoleného počtu vozidel daného úseku $i$. $R = rI$ penalizuje změnu107 časů zelených. Parametr $r$ ovlivňuje míru reakce systému a ja volen metodu pokus-oprava.108 Minimalizací tohoto kritéria pomcí \ref{eq_riccati} získáme zpětnovazebnou matici $L$,109 která určuje $g(t)$. Z rovnic \ref{eq_lq_feedback} a \ref{eq_delta_g} dostaneme výsledný vztah110 \begin{equation}\label{eq_tuc_feedback}111 g(t) = g^n - L x(t)112 \end{equation}. Toto řešení předpokládá, že známe hodnotu $g^n$, při které systém113 zůstává ve stavu $x^n$. Tak tomu ale většinou není. Při absenci znalosti $g^n$114 podle \cite{6_tuc_lq} odečteme $g(t) - g(t-1)$ a rovnice \ref{eq_tuc_feedback} nabývá tvaru115 \begin{equation}\label{eq_tuc_feedback_2}116 g(t) = g(t-1) - L( x(t) - x(t-1) )117 \end{equation}.118 36 119 37 … … 130 48 131 49 50 -
applications/doprava/texty/novotny_vyzk_LQ/Reinforcement_learning.tex
r1424 r1427 1 \section{Zpětnovazebné učení} 1 2 2 3 \section{Markovův rozhodvací proces} 3 \subsection{Markovův rozhodvací proces} 4 4 Markovův rozhodvací proces je alternativní metoda sloužící 5 5 k volbě strategií odhadem zisků z nich plynoucích do budoucna. … … 29 29 30 30 31 \subs ubsection{Dynamické programování}\label{sec:dynamic_programming}31 \subsection{Dynamické programování}\label{sec:dynamic_programming} 32 32 33 33 %asi trochu poupravit podle \cite{tlc_using_sarsa} … … 123 123 124 124 125 \subs ection{Učení na základě modelu (Model-based learning)}\label{sec:model_based_learning}125 \subsubsection{Učení na základě modelu (Model-based learning)}\label{sec:model_based_learning} 126 126 V této metodě, popsané v \cite{3_i_traff_light_c}, se modeluje 127 127 prostředí funkcemi $P(i,a,j)$ a $R(i,a,j)$, které jsou definované v -
applications/doprava/texty/novotny_vyzk_LQ/Reinforcement_learning_pouziti/Reinforcement_learning_pouziti.tex
r1424 r1427 1 \subsection{Použití zpětnovazebného učení} 2 3 \subsubsection{Zpětnovazebné učení na základě modelu} 1 \section{Použití zpětnovazebného učení} 4 2 5 3 V \cite{3_i_traff_light_c} je popsána simulace používající … … 30 28 hodnot $Q$.\\ 31 29 32 \subsection{Zhodnocení}33 30 34 Metode popsaná v článku \cite{3_i_traff_light_c} používá ohodnocovací funkci35 založenou na parametrech jednotlivých vozidel. Výhodou oproti pojetí, kdy agent36 představuje pouze signální skupinu, jsou například v tom, že není potřeba37 odhadovat délku fronty a úloha se celá zjednoduší. Například v publikaci38 \cite{tlc_using_sarsa} se musí používat k odhadu funkcí $V$ a $Q$ neuronová síť.39 Navíc tento systém umožňuje i výběr optimální cesty vozidla pro průjezd dopravní sítí.40 Nevýhodou tohoto pojetí je ovšem značná neuniverzálnost. Už pro počítačové testování41 tato metode znesnadnuje či úplně znemožňuje použít celou řadu dopravních simulátorů,42 které jsou pro simulaci po dlouhou dobu optimalizovány a43 jejichž nasazení značně zjednodušuje práci a urychluje vývoj.44 Navíc pokud je použit řadič, který obstarává logiku přepínání průjezdnosti a45 lze nastavovat pouze vnější parametry jako jsou délka cyklu a offset, je46 metoda, která potřebuje okamžitou změnu signalizace naprosto nevhodná,47 proto je toto řešení pro reálné nasazení v dnešní době obtížně použitelné.48 Zapojení některých myšlenek z článku \cite{3_i_traff_light_c} nebo použití49 zpětnovazevného učení k řešení dílčích problémů by však mohlo přinést zlepšení50 i do způsobu žešení popsaných v dalších kapitolách.51 31 -
applications/doprava/texty/novotny_vyzk_LQ/vyzk.aux
r1425 r1427 27 27 \citation{wooldridge} 28 28 \citation{wooldridge} 29 \@writefile{toc}{\contentsline {chapter}{\numberline {2}M ultiagentn\IeC {\'\i } syst\IeC {\'e}my}{3}{chapter.2}}30 \@writefile{lof}{\addvspace {10\p@ }} 31 \@writefile{lot}{\addvspace {10\p@ }} 32 \@writefile{toc}{\contentsline {section}{\numberline {2.1} \IeC {\'U}vod}{3}{section.2.1}}29 \@writefile{toc}{\contentsline {chapter}{\numberline {2}Matematick\IeC {\'e} metody rozhodov\IeC {\'a}n\IeC {\'\i }}{3}{chapter.2}} 30 \@writefile{lof}{\addvspace {10\p@ }} 31 \@writefile{lot}{\addvspace {10\p@ }} 32 \@writefile{toc}{\contentsline {section}{\numberline {2.1}Multiagentn\IeC {\'\i } syst\IeC {\'e}my}{3}{section.2.1}} 33 33 \@writefile{toc}{\contentsline {subsection}{\numberline {2.1.1}Historie}{3}{subsection.2.1.1}} 34 34 \@writefile{toc}{\contentsline {subsection}{\numberline {2.1.2}Agent}{3}{subsection.2.1.2}} 35 35 \@writefile{thm}{\contentsline {definition}{{Definice}{2.{1}}{Agent}}{3}{definition.2.1}} 36 36 \newlabel{de:agent01}{{2.{1}}{3}{Agent\relax }{definition.2.1}{}} 37 \@writefile{toc}{\contentsline {section}{\numberline {2.2}Druhy prost\IeC {\v r}ed\IeC {\'\i }}{4}{section.2.2}} 38 \@writefile{toc}{\contentsline {section}{\numberline {2.3}Interakce agent\IeC {\r u}}{4}{section.2.3}} 39 \@writefile{toc}{\contentsline {subsection}{\numberline {2.3.1}Stavy prost\IeC {\v r}ed\IeC {\'\i } a preference agent\IeC {\r u}}{4}{subsection.2.3.1}} 37 \@writefile{toc}{\contentsline {subsection}{\numberline {2.1.3}Druhy prost\IeC {\v r}ed\IeC {\'\i }}{4}{subsection.2.1.3}} 38 \@writefile{toc}{\contentsline {subsection}{\numberline {2.1.4}Stavy prost\IeC {\v r}ed\IeC {\'\i } a preference agent\IeC {\r u}}{4}{subsection.2.1.4}} 40 39 \@writefile{thm}{\contentsline {definition}{{Definice}{2.{2}}{Uspořádání na množině všech stavů}}{4}{definition.2.2}} 40 \citation{wooldridge} 41 41 \global\def\markiidefinition{\ensuremath {\blacktriangleright }} 42 \@writefile{toc}{\contentsline {chapter}{\numberline {3}V\IeC {\'y}b\IeC {\v e}r strategie genta}{6}{chapter.3}} 43 \@writefile{lof}{\addvspace {10\p@ }} 44 \@writefile{lot}{\addvspace {10\p@ }} 45 \@writefile{toc}{\contentsline {section}{\numberline {3.1}V\IeC {\'y}b\IeC {\v e}r strategie podle teorie her}{6}{section.3.1}} 46 \@writefile{thm}{\contentsline {definition}{{Definice}{3.{1}}{Dominance množiny}}{6}{definition.3.1}} 42 \@writefile{toc}{\contentsline {section}{\numberline {2.2}V\IeC {\'y}b\IeC {\v e}r strategie podle teorie her}{5}{section.2.2}} 43 \@writefile{thm}{\contentsline {definition}{{Definice}{2.{3}}{Dominance množiny}}{5}{definition.2.3}} 47 44 \citation{3_i_traff_light_c} 48 45 \global\def\markiiidefinition{\ensuremath {\blacktriangleright }} 49 \@writefile{thm}{\contentsline {definition}{{Definice}{ 3.{2}}{Množina výsledků}}{7}{definition.3.2}}46 \@writefile{thm}{\contentsline {definition}{{Definice}{2.{4}}{Množina výsledků}}{6}{definition.2.4}} 50 47 \global\def\markivdefinition{\ensuremath {\blacktriangleright }} 51 \@writefile{thm}{\contentsline {definition}{{Definice}{3.{3}}{Dominance strategie}}{7}{definition.3.3}} 52 \@writefile{thm}{\contentsline {definition}{{Definice}{3.{4}}{Nashova rovnost}}{7}{definition.3.4}} 53 \newlabel{de:nash_equlibrium}{{3.{4}}{7}{Výběr strategie podle teorie her\relax }{definition.3.4}{}} 54 \@writefile{toc}{\contentsline {section}{\numberline {3.2}Markov\IeC {\r u}v rozhodvac\IeC {\'\i } proces}{7}{section.3.2}} 55 \@writefile{thm}{\contentsline {definition}{{Definice}{3.{5}}{Markovův rozhodovací proces}}{7}{definition.3.5}} 56 \newlabel{de:markov_decision_process}{{3.{5}}{7}{Markovův rozhodvací proces\relax }{definition.3.5}{}} 57 \citation{3_i_traff_light_c} 58 \citation{3_i_traff_light_c} 48 \@writefile{thm}{\contentsline {definition}{{Definice}{2.{5}}{Dominance strategie}}{6}{definition.2.5}} 49 \@writefile{thm}{\contentsline {definition}{{Definice}{2.{6}}{Nashova rovnost}}{6}{definition.2.6}} 50 \newlabel{de:nash_equlibrium}{{2.{6}}{6}{Výběr strategie podle teorie her\relax }{definition.2.6}{}} 51 \@writefile{toc}{\contentsline {section}{\numberline {2.3}Zp\IeC {\v e}tnovazebn\IeC {\'e} u\IeC {\v c}en\IeC {\'\i }}{6}{section.2.3}} 52 \@writefile{toc}{\contentsline {subsection}{\numberline {2.3.1}Markov\IeC {\r u}v rozhodvac\IeC {\'\i } proces}{6}{subsection.2.3.1}} 53 \citation{3_i_traff_light_c} 54 \citation{3_i_traff_light_c} 55 \@writefile{thm}{\contentsline {definition}{{Definice}{2.{7}}{Markovův rozhodovací proces}}{7}{definition.2.7}} 56 \newlabel{de:markov_decision_process}{{2.{7}}{7}{Markovův rozhodvací proces\relax }{definition.2.7}{}} 57 \global\def\markviidefinitioni{\ensuremath {\blacktriangleright }} 58 \@writefile{toc}{\contentsline {subsection}{\numberline {2.3.2}Dynamick\IeC {\'e} programov\IeC {\'a}n\IeC {\'\i }}{7}{subsection.2.3.2}} 59 \newlabel{sec:dynamic_programming}{{2.3.2}{7}{Dynamické programování\relax }{subsection.2.3.2}{}} 60 \@writefile{thm}{\contentsline {definition}{{Definice}{2.{8}}{V-funkce}}{7}{definition.2.8}} 61 \newlabel{de:v_function}{{2.{8}}{7}{Dynamické programování\relax }{definition.2.8}{}} 62 \global\def\markviiidefinitioni{\ensuremath {\blacktriangleright }} 59 63 \citation{dynamic_programming} 60 \global\def\markviidefinitioni{\ensuremath {\blacktriangleright }} 61 \@writefile{toc}{\contentsline {subsubsection}{\numberline {3.2.0.1}Dynamick\IeC {\'e} programov\IeC {\'a}n\IeC {\'\i }}{8}{subsubsection.3.2.0.1}} 62 \newlabel{sec:dynamic_programming}{{3.2.0.1}{8}{Dynamické programování\relax }{subsubsection.3.2.0.1}{}} 63 \@writefile{thm}{\contentsline {definition}{{Definice}{3.{6}}{V-funkce}}{8}{definition.3.6}} 64 \newlabel{de:v_function}{{3.{6}}{8}{Dynamické programování\relax }{definition.3.6}{}} 65 \global\def\markviiidefinitioni{\ensuremath {\blacktriangleright }} 66 \@writefile{thm}{\contentsline {definition}{{Definice}{3.{7}}{Q-function}}{8}{definition.3.7}} 67 \newlabel{de:q_function}{{3.{7}}{8}{Dynamické programování\relax }{definition.3.7}{}} 68 \citation{3_i_traff_light_c} 69 \citation{3_i_traff_light_c} 64 \citation{3_i_traff_light_c} 65 \citation{3_i_traff_light_c} 66 \@writefile{thm}{\contentsline {definition}{{Definice}{2.{9}}{Q-function}}{8}{definition.2.9}} 67 \newlabel{de:q_function}{{2.{9}}{8}{Dynamické programování\relax }{definition.2.9}{}} 68 \global\def\markixdefinition{\ensuremath {\blacktriangleright }} 69 \@writefile{thm}{\contentsline {definition}{{Definice}{2.{10}}{Bellmanova rovnice optimality}}{8}{definition.2.10}} 70 \newlabel{de:bellman_equation_of_optimality}{{2.{10}}{8}{Dynamické programování\relax }{definition.2.10}{}} 71 \global\def\markxdefinition{\ensuremath {\blacktriangleright }} 70 72 \citation{q_learning} 71 73 \citation{learning_to_predict} 72 \global\def\markixdefinition{\ensuremath {\blacktriangleright }} 73 \@writefile{thm}{\contentsline {definition}{{Definice}{3.{8}}{Bellmanova rovnice optimality}}{9}{definition.3.8}} 74 \newlabel{de:bellman_equation_of_optimality}{{3.{8}}{9}{Dynamické programování\relax }{definition.3.8}{}} 75 \global\def\markxdefinition{\ensuremath {\blacktriangleright }} 76 \@writefile{toc}{\contentsline {subsection}{\numberline {3.2.1}Zp\IeC {\v e}tnovazebn\IeC {\'e} u\IeC {\v c}en\IeC {\'\i } (Reinforcement learning)}{9}{subsection.3.2.1}} 77 \citation{3_i_traff_light_c} 78 \citation{3_i_traff_light_c} 79 \citation{3_i_traff_light_c} 80 \citation{3_i_traff_light_c} 81 \@writefile{toc}{\contentsline {subsubsection}{\numberline {3.2.1.1}Q-u\IeC {\v c}en\IeC {\'\i } (Q-learning)}{10}{subsubsection.3.2.1.1}} 82 \@writefile{toc}{\contentsline {subsection}{\numberline {3.2.2}U\IeC {\v c}en\IeC {\'\i } na z\IeC {\'a}klad\IeC {\v e} modelu (Model-based learning)}{10}{subsection.3.2.2}} 83 \newlabel{sec:model_based_learning}{{3.2.2}{10}{Učení na základě modelu (Model-based learning)\relax }{subsection.3.2.2}{}} 84 \@writefile{thm}{\contentsline {definition}{{Definice}{3.{9}}{MLM}}{10}{definition.3.9}} 85 \newlabel{de:mlm}{{3.{9}}{10}{Učení na základě modelu (Model-based learning)\relax }{definition.3.9}{}} 74 \citation{3_i_traff_light_c} 75 \citation{3_i_traff_light_c} 76 \citation{3_i_traff_light_c} 77 \@writefile{toc}{\contentsline {subsection}{\numberline {2.3.3}Zp\IeC {\v e}tnovazebn\IeC {\'e} u\IeC {\v c}en\IeC {\'\i } (Reinforcement learning)}{9}{subsection.2.3.3}} 78 \@writefile{toc}{\contentsline {subsubsection}{\numberline {2.3.3.1}Q-u\IeC {\v c}en\IeC {\'\i } (Q-learning)}{9}{subsubsection.2.3.3.1}} 79 \@writefile{toc}{\contentsline {subsubsection}{\numberline {2.3.3.2}U\IeC {\v c}en\IeC {\'\i } na z\IeC {\'a}klad\IeC {\v e} modelu (Model-based learning)}{9}{subsubsection.2.3.3.2}} 80 \newlabel{sec:model_based_learning}{{2.3.3.2}{9}{Učení na základě modelu (Model-based learning)\relax }{subsubsection.2.3.3.2}{}} 81 \@writefile{thm}{\contentsline {definition}{{Definice}{2.{11}}{MLM}}{9}{definition.2.11}} 82 \newlabel{de:mlm}{{2.{11}}{9}{Učení na základě modelu (Model-based learning)\relax }{definition.2.11}{}} 83 \citation{4_rmm_formalization} 84 \citation{4_rmm_formalization} 85 \citation{4_rmm_formalization} 86 86 \global\def\markxidefinition{\ensuremath {\blacktriangleright }} 87 \citation{3_i_traff_light_c} 88 \citation{3_i_traff_light_c} 89 \citation{tlc_using_sarsa} 90 \citation{3_i_traff_light_c} 91 \@writefile{toc}{\contentsline {subsection}{\numberline {3.2.3}Pou\IeC {\v z}it\IeC {\'\i } zp\IeC {\v e}tnovazebn\IeC {\'e}ho u\IeC {\v c}en\IeC {\'\i }}{11}{subsection.3.2.3}} 92 \@writefile{toc}{\contentsline {subsubsection}{\numberline {3.2.3.1}Zp\IeC {\v e}tnovazebn\IeC {\'e} u\IeC {\v c}en\IeC {\'\i } na z\IeC {\'a}klad\IeC {\v e} modelu}{11}{subsubsection.3.2.3.1}} 93 \@writefile{toc}{\contentsline {subsection}{\numberline {3.2.4}Zhodnocen\IeC {\'\i }}{11}{subsection.3.2.4}} 94 \citation{4_rmm_formalization} 95 \citation{4_rmm_formalization} 96 \@writefile{toc}{\contentsline {section}{\numberline {3.3}RMM - Rekurzivn\IeC {\'\i } modelov\IeC {\'e} metody}{12}{section.3.3}} 97 \@writefile{toc}{\contentsline {subsection}{\numberline {3.3.1}Form\IeC {\'a}ln\IeC {\'\i } definice}{12}{subsection.3.3.1}} 98 \@writefile{thm}{\contentsline {definition}{{Definice}{3.{10}}{Matice zisků}}{12}{definition.3.10}} 99 \newlabel{de:payoff_matrix}{{3.{10}}{12}{Formální definice\relax }{definition.3.10}{}} 100 \citation{4_rmm_formalization} 101 \citation{4_rmm_formalization} 102 \citation{4_rmm_formalization} 103 \citation{4_rmm_formalization} 104 \@writefile{thm}{\contentsline {definition}{{Definice}{3.{11}}{Rekursivní modelová struktura}}{13}{definition.3.11}} 105 \newlabel{de:rms}{{3.{11}}{13}{Formální definice\relax }{definition.3.11}{}} 106 \@writefile{thm}{\contentsline {definition}{{Definice}{3.{12}}{Rekursivní model}}{13}{definition.3.12}} 107 \newlabel{de:rm}{{3.{12}}{13}{Formální definice\relax }{definition.3.12}{}} 108 \citation{4_rmm_formalization} 109 \@writefile{toc}{\contentsline {subsection}{\numberline {3.3.2}Rozhodovac\IeC {\'\i } algoritmus}{14}{subsection.3.3.2}} 110 \@writefile{thm}{\contentsline {definition}{{Definice}{3.{13}}{Užitečnost}}{14}{definition.3.13}} 111 \newlabel{de:utility}{{3.{13}}{14}{Rozhodovací algoritmus\relax }{definition.3.13}{}} 112 \citation{4_rmm_formalization} 87 \@writefile{toc}{\contentsline {section}{\numberline {2.4}RMM - Rekurzivn\IeC {\'\i } modelov\IeC {\'e} metody}{10}{section.2.4}} 88 \@writefile{toc}{\contentsline {subsection}{\numberline {2.4.1}Form\IeC {\'a}ln\IeC {\'\i } definice}{10}{subsection.2.4.1}} 89 \@writefile{thm}{\contentsline {definition}{{Definice}{2.{12}}{Matice zisků}}{10}{definition.2.12}} 90 \newlabel{de:payoff_matrix}{{2.{12}}{10}{Formální definice\relax }{definition.2.12}{}} 91 \citation{4_rmm_formalization} 92 \citation{4_rmm_formalization} 93 \citation{4_rmm_formalization} 94 \@writefile{thm}{\contentsline {definition}{{Definice}{2.{13}}{Rekursivní modelová struktura}}{11}{definition.2.13}} 95 \newlabel{de:rms}{{2.{13}}{11}{Formální definice\relax }{definition.2.13}{}} 96 \@writefile{thm}{\contentsline {definition}{{Definice}{2.{14}}{Rekursivní model}}{11}{definition.2.14}} 97 \newlabel{de:rm}{{2.{14}}{11}{Formální definice\relax }{definition.2.14}{}} 98 \citation{4_rmm_formalization} 99 \citation{4_rmm_formalization} 100 \@writefile{toc}{\contentsline {subsection}{\numberline {2.4.2}Rozhodovac\IeC {\'\i } algoritmus}{12}{subsection.2.4.2}} 101 \@writefile{thm}{\contentsline {definition}{{Definice}{2.{15}}{Užitečnost}}{12}{definition.2.15}} 102 \newlabel{de:utility}{{2.{15}}{12}{Rozhodovací algoritmus\relax }{definition.2.15}{}} 103 \global\def\markxvdefinition{\ensuremath {\blacktriangleright }} 113 104 \citation{5_bayes_learn} 114 105 \citation{5_bayes_learn} 115 \ global\def\markxvdefinition{\ensuremath {\blacktriangleright }}116 \@writefile{toc}{\contentsline {s ubsection}{\numberline {3.3.3}Bayesovsk\IeC {\'e} u\IeC {\v c}en\IeC {\'\i }}{15}{subsection.3.3.3}}117 \newlabel{sec:bayes}{{ 3.3.3}{15}{Bayesovské učení\relax }{subsection.3.3.3}{}}118 \@writefile{thm}{\contentsline {definition}{{Definice}{ 3.{14}}{Podmíněná pravděpodobnost}}{15}{definition.3.14}}119 \newlabel{de:podm_pravd}{{ 3.{14}}{15}{Bayesovské učení\relax }{definition.3.14}{}}106 \citation{5_bayes_learn} 107 \@writefile{toc}{\contentsline {section}{\numberline {2.5}Bayesovsk\IeC {\'e} u\IeC {\v c}en\IeC {\'\i }}{13}{section.2.5}} 108 \newlabel{sec:bayes}{{2.5}{13}{Bayesovské učení\relax }{section.2.5}{}} 109 \@writefile{thm}{\contentsline {definition}{{Definice}{2.{16}}{Podmíněná pravděpodobnost}}{13}{definition.2.16}} 110 \newlabel{de:podm_pravd}{{2.{16}}{13}{Bayesovské učení\relax }{definition.2.16}{}} 120 111 \global\def\markxvidefinition{\ensuremath {\blacktriangleright }} 121 \@writefile{thm}{\contentsline {proposition}{{Věta}{3.{1}}{Bayesova věta}}{15}{proposition.3.1}} 122 \newlabel{v:bayes}{{3.{1}}{15}{Bayesovské učení\relax }{proposition.3.1}{}} 123 \citation{5_bayes_learn} 112 \@writefile{thm}{\contentsline {proposition}{{Věta}{2.{1}}{Bayesova věta}}{13}{proposition.2.1}} 113 \newlabel{v:bayes}{{2.{1}}{13}{Bayesovské učení\relax }{proposition.2.1}{}} 114 \@writefile{toc}{\contentsline {subsubsection}{\numberline {2.5.0.1}V\IeC {\v e}rohodnostn\IeC {\'\i } funkce}{13}{subsubsection.2.5.0.1}} 115 \citation{6_tuc_lq} 116 \@writefile{thm}{\contentsline {definition}{{Definice}{2.{17}}{Věrohodnostní funkce}}{14}{definition.2.17}} 117 \newlabel{de:ver_fce}{{2.{17}}{14}{Věrohodnostní funkce\relax }{definition.2.17}{}} 118 \global\def\markxviidefinition{\ensuremath {\blacktriangleright }} 119 \@writefile{toc}{\contentsline {section}{\numberline {2.6}LQ \IeC {\v r}\IeC {\'\i }zen\IeC {\'\i }}{14}{section.2.6}} 120 \citation{6_tuc_lq} 121 \citation{7_lq_methods} 122 \newlabel{eq_lq_feedback}{{2.2}{15}{LQ řízení\relax }{equation.2.6.2}{}} 123 \newlabel{eq_quadratic_criterion}{{2.3}{15}{LQ řízení\relax }{equation.2.6.3}{}} 124 \newlabel{eq_riccati}{{2.4}{15}{LQ řízení\relax }{equation.2.6.4}{}} 125 \newlabel{eq_riccati_2}{{2.5}{15}{LQ řízení\relax }{equation.2.6.5}{}} 126 \citation{3_i_traff_light_c} 127 \citation{3_i_traff_light_c} 128 \@writefile{toc}{\contentsline {chapter}{\numberline {3}Pou\IeC {\v z}it\IeC {\'\i } rozhodovac\IeC {\'\i }ch metod v \IeC {\v r}\IeC {\'\i }zen\IeC {\'\i } dopravy}{16}{chapter.3}} 129 \@writefile{lof}{\addvspace {10\p@ }} 130 \@writefile{lot}{\addvspace {10\p@ }} 131 \@writefile{toc}{\contentsline {section}{\numberline {3.1}Pou\IeC {\v z}it\IeC {\'\i } zp\IeC {\v e}tnovazebn\IeC {\'e}ho u\IeC {\v c}en\IeC {\'\i }}{16}{section.3.1}} 124 132 \citation{1_rmm_bayes_learning} 125 \@writefile{toc}{\contentsline {subsubsection}{\numberline {3.3.3.1}V\IeC {\v e}rohodnostn\IeC {\'\i } funkce}{16}{subsubsection.3.3.3.1}}126 \@writefile{thm}{\contentsline {definition}{{Definice}{3.{15}}{Věrohodnostní funkce}}{16}{definition.3.15}}127 \newlabel{de:ver_fce}{{3.{15}}{16}{Věrohodnostní funkce\relax }{definition.3.15}{}}128 \global\def\markxviidefinition{\ensuremath {\blacktriangleright }}129 \@writefile{toc}{\contentsline {subsection}{\numberline {3.3.4}Pou\IeC {\v z}it\IeC {\'\i } RMM a Bayesova u\IeC {\v c}en\IeC {\'\i } v decentralizovan\IeC {\'e}m \IeC {\v r}\IeC {\'\i }zen\IeC {\'\i } dopravy}{16}{subsection.3.3.4}}130 133 \citation{1_rmm_bayes_learning} 131 134 \citation{4_rmm_formalization} 132 135 \citation{1_rmm_bayes_learning} 133 \@writefile{lof}{\contentsline {figure}{\numberline {3.1}{\ignorespaces T\IeC {\v r}\IeC {\'\i }\IeC {\'u}rov\IeC {\v n}ov\IeC {\'a} rekurzivn\IeC {\'\i } modelov\IeC {\'a} struktura agenta $R_1$\relax }}{17}{figure.caption.3}} 136 \@writefile{toc}{\contentsline {section}{\numberline {3.2}Pou\IeC {\v z}it\IeC {\'\i } RMM a Bayesova u\IeC {\v c}en\IeC {\'\i } v decentralizovan\IeC {\'e}m \IeC {\v r}\IeC {\'\i }zen\IeC {\'\i } dopravy}{17}{section.3.2}} 137 \@writefile{lof}{\contentsline {figure}{\numberline {3.1}{\ignorespaces T\IeC {\v r}\IeC {\'\i }\IeC {\'u}rov\IeC {\v n}ov\IeC {\'a} rekurzivn\IeC {\'\i } modelov\IeC {\'a} struktura agenta $R_1$\relax }}{18}{figure.caption.3}} 134 138 \providecommand*\caption@xref[2]{\@setref\relax\@undefined{#1}} 135 \newlabel{fig:rmm_structure}{{3.1}{17}{Tříúrovňová rekurzivní modelová struktura agenta $R_1$\relax \relax }{figure.caption.3}{}} 136 \citation{4_rmm_formalization} 137 \citation{6_tuc_lq} 138 \@writefile{lof}{\contentsline {figure}{\numberline {3.2}{\ignorespaces V\IeC {\'y}sledky m\IeC {\v e}\IeC {\v r}en\IeC {\'\i }. M1 - RMM, M2 - pevn\IeC {\'e} cykly, M3 - "Hill-climbing" \relax }}{18}{figure.caption.4}} 139 \newlabel{fig:rmm_results}{{3.2}{18}{Výsledky měření. M1 - RMM, M2 - pevné cykly, M3 - "Hill-climbing" \relax \relax }{figure.caption.4}{}} 140 \@writefile{toc}{\contentsline {subsection}{\numberline {3.3.5}Zhodnocen\IeC {\'\i }}{18}{subsection.3.3.5}} 141 \@writefile{toc}{\contentsline {section}{\numberline {3.4}LQ \IeC {\v r}\IeC {\'\i }zen\IeC {\'\i }}{18}{section.3.4}} 142 \citation{6_tuc_lq} 143 \citation{7_lq_methods} 144 \citation{6_tuc_lq} 145 \newlabel{eq_lq_feedback}{{3.2}{19}{LQ řízení\relax }{equation.3.4.2}{}} 146 \newlabel{eq_quadratic_criterion}{{3.3}{19}{LQ řízení\relax }{equation.3.4.3}{}} 147 \newlabel{eq_riccati}{{3.4}{19}{LQ řízení\relax }{equation.3.4.4}{}} 148 \newlabel{eq_riccati_2}{{3.5}{19}{LQ řízení\relax }{equation.3.4.5}{}} 149 \@writefile{toc}{\contentsline {subsection}{\numberline {3.4.1}Pou\IeC {\v z}it\IeC {\'\i } LQ \IeC {\v r}\IeC {\'\i }zen\IeC {\'\i } ve strategii TUC}{19}{subsection.3.4.1}} 150 \newlabel{eq_tuc_1}{{3.6}{19}{Použití LQ řízení ve strategii TUC\relax }{equation.3.4.6}{}} 151 \citation{6_tuc_lq} 152 \newlabel{eq:tuc_u}{{3.9}{20}{Použití LQ řízení ve strategii TUC\relax }{equation.3.4.9}{}} 153 \newlabel{eq_tuc_2}{{3.10}{20}{Použití LQ řízení ve strategii TUC\relax }{equation.3.4.10}{}} 154 \newlabel{eq_tuc_nom}{{3.11}{20}{Použití LQ řízení ve strategii TUC\relax }{equation.3.4.11}{}} 155 \newlabel{eq_delta_g}{{3.12}{20}{Použití LQ řízení ve strategii TUC\relax }{equation.3.4.12}{}} 156 \newlabel{eq_tuc_3}{{3.13}{20}{Použití LQ řízení ve strategii TUC\relax }{equation.3.4.13}{}} 157 \newlabel{eq_tuc_4}{{3.14}{20}{Použití LQ řízení ve strategii TUC\relax }{equation.3.4.14}{}} 158 \citation{6_tuc_lq} 159 \citation{6_tuc_lq} 160 \@writefile{toc}{\contentsline {subsubsection}{\numberline {3.4.1.1}Kvadratick\IeC {\'e} krit\IeC {\'e}rium}{21}{subsubsection.3.4.1.1}} 161 \newlabel{eq_tuc_crit}{{3.15}{21}{Kvadratické kritérium\relax }{equation.3.4.15}{}} 162 \newlabel{eq_tuc_feedback}{{3.16}{21}{Kvadratické kritérium\relax }{equation.3.4.16}{}} 163 \newlabel{eq_tuc_feedback_2}{{3.17}{21}{Kvadratické kritérium\relax }{equation.3.4.17}{}} 164 \@writefile{toc}{\contentsline {chapter}{\numberline {4}Implementace}{22}{chapter.4}} 165 \@writefile{lof}{\addvspace {10\p@ }} 166 \@writefile{lot}{\addvspace {10\p@ }} 167 \@writefile{toc}{\contentsline {section}{\numberline {4.1}Pou\IeC {\v z}it\IeC {\'a} metoda}{24}{section.4.1}} 168 \@writefile{toc}{\contentsline {subsection}{\numberline {4.1.1}P\IeC {\v r}echodov\IeC {\'e} vztahy}{24}{subsection.4.1.1}} 169 \newlabel{eq:my_trans_01}{{4.1}{24}{Přechodové vztahy\relax }{equation.4.1.1}{}} 170 \newlabel{eq:my_trans_02}{{4.5}{24}{Přechodové vztahy\relax }{equation.4.1.5}{}} 171 \newlabel{eq:my_trans_mat}{{4.7}{25}{Přechodové vztahy\relax }{equation.4.1.7}{}} 172 \@writefile{toc}{\contentsline {subsection}{\numberline {4.1.2}Minimalizace krit\IeC {\'e}ria}{25}{subsection.4.1.2}} 173 \newlabel{eq:prechod_subs_01}{{4.9}{25}{Minimalizace kritéria\relax }{equation.4.1.9}{}} 174 \newlabel{eq:prechod_mat_po_subs}{{4.11}{25}{Minimalizace kritéria\relax }{equation.4.1.11}{}} 175 \newlabel{eq:J}{{4.12}{25}{Minimalizace kritéria\relax }{equation.4.1.12}{}} 176 \newlabel{eq:J_sloz}{{4.15}{26}{Minimalizace kritéria\relax }{equation.4.1.15}{}} 177 \@writefile{toc}{\contentsline {subsubsection}{\numberline {4.1.2.1}Implementace minimalizace}{27}{subsubsection.4.1.2.1}} 139 \newlabel{fig:rmm_structure}{{3.1}{18}{Tříúrovňová rekurzivní modelová struktura agenta $R_1$\relax \relax }{figure.caption.3}{}} 140 \citation{3_i_traff_light_c} 141 \citation{tlc_using_sarsa} 142 \citation{3_i_traff_light_c} 143 \@writefile{lof}{\contentsline {figure}{\numberline {3.2}{\ignorespaces V\IeC {\'y}sledky m\IeC {\v e}\IeC {\v r}en\IeC {\'\i }. M1 - RMM, M2 - pevn\IeC {\'e} cykly, M3 - "Hill-climbing" \relax }}{19}{figure.caption.4}} 144 \newlabel{fig:rmm_results}{{3.2}{19}{Výsledky měření. M1 - RMM, M2 - pevné cykly, M3 - "Hill-climbing" \relax \relax }{figure.caption.4}{}} 145 \@writefile{toc}{\contentsline {section}{\numberline {3.3}Zhodnocen\IeC {\'\i }}{19}{section.3.3}} 146 \@writefile{toc}{\contentsline {subsection}{\numberline {3.3.1}Pou\IeC {\v z}it\IeC {\'\i } zp\IeC {\v e}tnovazebn\IeC {\'e}ho u\IeC {\v c}en\IeC {\'\i }}{19}{subsection.3.3.1}} 147 \citation{4_rmm_formalization} 148 \citation{6_tuc_lq} 149 \@writefile{toc}{\contentsline {subsection}{\numberline {3.3.2}Pou\IeC {\v z}it\IeC {\'\i } RMM a Bayesova u\IeC {\v c}en\IeC {\'\i }}{20}{subsection.3.3.2}} 150 \@writefile{toc}{\contentsline {subsection}{\numberline {3.3.3}Pou\IeC {\v z}it\IeC {\'\i } LQ \IeC {\v r}\IeC {\'\i }zen\IeC {\'\i } ve strategii TUC}{20}{subsection.3.3.3}} 151 \newlabel{eq_tuc_1}{{3.1}{20}{Použití LQ řízení ve strategii TUC\relax }{equation.3.3.1}{}} 152 \citation{6_tuc_lq} 153 \citation{6_tuc_lq} 154 \newlabel{eq:tuc_u}{{3.4}{21}{Použití LQ řízení ve strategii TUC\relax }{equation.3.3.4}{}} 155 \newlabel{eq_tuc_2}{{3.5}{21}{Použití LQ řízení ve strategii TUC\relax }{equation.3.3.5}{}} 156 \newlabel{eq_tuc_nom}{{3.6}{21}{Použití LQ řízení ve strategii TUC\relax }{equation.3.3.6}{}} 157 \newlabel{eq_delta_g}{{3.7}{21}{Použití LQ řízení ve strategii TUC\relax }{equation.3.3.7}{}} 158 \newlabel{eq_tuc_3}{{3.8}{21}{Použití LQ řízení ve strategii TUC\relax }{equation.3.3.8}{}} 159 \newlabel{eq_tuc_4}{{3.9}{21}{Použití LQ řízení ve strategii TUC\relax }{equation.3.3.9}{}} 160 \@writefile{toc}{\contentsline {subsubsection}{\numberline {3.3.3.1}Kvadratick\IeC {\'e} krit\IeC {\'e}rium}{21}{subsubsection.3.3.3.1}} 161 \newlabel{eq_tuc_crit}{{3.10}{21}{Kvadratické kritérium\relax }{equation.3.3.10}{}} 162 \citation{6_tuc_lq} 163 \newlabel{eq_tuc_feedback}{{3.11}{22}{Kvadratické kritérium\relax }{equation.3.3.11}{}} 164 \newlabel{eq_tuc_feedback_2}{{3.12}{22}{Kvadratické kritérium\relax }{equation.3.3.12}{}} 165 \@writefile{toc}{\contentsline {chapter}{\numberline {4}Implementace}{23}{chapter.4}} 166 \@writefile{lof}{\addvspace {10\p@ }} 167 \@writefile{lot}{\addvspace {10\p@ }} 168 \@writefile{toc}{\contentsline {section}{\numberline {4.1}Pou\IeC {\v z}it\IeC {\'a} metoda}{25}{section.4.1}} 169 \@writefile{toc}{\contentsline {subsection}{\numberline {4.1.1}P\IeC {\v r}echodov\IeC {\'e} vztahy}{25}{subsection.4.1.1}} 170 \newlabel{eq:my_trans_01}{{4.1}{25}{Přechodové vztahy\relax }{equation.4.1.1}{}} 171 \newlabel{eq:my_trans_02}{{4.5}{25}{Přechodové vztahy\relax }{equation.4.1.5}{}} 172 \newlabel{eq:my_trans_mat}{{4.7}{26}{Přechodové vztahy\relax }{equation.4.1.7}{}} 173 \@writefile{toc}{\contentsline {subsection}{\numberline {4.1.2}Minimalizace krit\IeC {\'e}ria}{26}{subsection.4.1.2}} 174 \newlabel{eq:prechod_subs_01}{{4.9}{26}{Minimalizace kritéria\relax }{equation.4.1.9}{}} 175 \newlabel{eq:prechod_mat_po_subs}{{4.11}{26}{Minimalizace kritéria\relax }{equation.4.1.11}{}} 176 \newlabel{eq:J}{{4.12}{26}{Minimalizace kritéria\relax }{equation.4.1.12}{}} 177 \newlabel{eq:J_sloz}{{4.15}{27}{Minimalizace kritéria\relax }{equation.4.1.15}{}} 178 \@writefile{toc}{\contentsline {subsubsection}{\numberline {4.1.2.1}Implementace minimalizace}{28}{subsubsection.4.1.2.1}} 178 179 \citation{aimsunget} 179 \@writefile{toc}{\contentsline {section}{\numberline {4.2}Simulace}{2 8}{section.4.2}}180 \@writefile{toc}{\contentsline {subsubsection}{\numberline {4.2.0.2}VGS API}{ 29}{subsubsection.4.2.0.2}}181 \newlabel{ss:vgs_api}{{4.2.0.2}{ 29}{VGS API\relax }{subsubsection.4.2.0.2}{}}182 \@writefile{toc}{\contentsline {subsection}{\numberline {4.2.1}\IeC {\v R}adi\IeC {\v c}e}{3 0}{subsection.4.2.1}}183 \@writefile{toc}{\contentsline {subsection}{\numberline {4.2.2}Oblast simulace}{3 0}{subsection.4.2.2}}184 \newlabel{ss:oblast_simulace}{{4.2.2}{3 0}{Oblast simulace\relax }{subsection.4.2.2}{}}185 \@writefile{lof}{\contentsline {figure}{\numberline {4.1}{\ignorespaces K\IeC {\v r}i\IeC {\v z}ovatka 601\relax }}{3 0}{figure.caption.5}}186 \newlabel{fig:601}{{4.1}{3 0}{Křižovatka 601\relax \relax }{figure.caption.5}{}}187 \citation{6_tuc_lq} 188 \@writefile{lof}{\contentsline {figure}{\numberline {4.2}{\ignorespaces K\IeC {\v r}i\IeC {\v z}ovatka 495\relax }}{3 1}{figure.caption.6}}189 \newlabel{fig:601}{{4.2}{3 1}{Křižovatka 495\relax \relax }{figure.caption.6}{}}190 \@writefile{toc}{\contentsline {section}{\numberline {4.3}Popis algoritmu}{3 1}{section.4.3}}191 \@writefile{toc}{\contentsline {section}{\numberline {4.4}Mo\IeC {\v z}n\IeC {\'e} vylep\IeC {\v s}en\IeC {\'\i } do budoucna}{3 2}{section.4.4}}192 \@writefile{toc}{\contentsline {subsubsection}{\numberline {4.4.0.1}Model toku}{3 2}{subsubsection.4.4.0.1}}193 \newlabel{eq:teor_tok}{{4.27}{3 2}{Model toku\relax }{equation.4.4.27}{}}194 \newlabel{eq:exp_tok}{{4.28}{3 2}{Model toku\relax }{equation.4.4.28}{}}195 \newlabel{eq:lin_tok}{{4.31}{3 2}{Model toku\relax }{equation.4.4.31}{}}196 \@writefile{toc}{\contentsline {subsection}{\numberline {4.4.1}Odhdad odbo\IeC {\v c}ovac\IeC {\'\i }ch pom\IeC {\v e}r\IeC {\r u}}{3 2}{subsection.4.4.1}}197 \@writefile{toc}{\contentsline {chapter}{\numberline {5}V\IeC {\'y}sledky}{3 4}{chapter.5}}198 \@writefile{lof}{\addvspace {10\p@ }} 199 \@writefile{lot}{\addvspace {10\p@ }} 200 \@writefile{toc}{\contentsline {section}{\numberline {5.1}Sc\IeC {\'e}n\IeC {\'a}\IeC {\v r} 1}{3 4}{section.5.1}}201 \@writefile{lof}{\contentsline {figure}{\numberline {5.1}{\ignorespaces Pr\IeC {\r u}b\IeC {\v e}h d\IeC {\'e}lky cyklu\relax }}{3 5}{figure.caption.7}}202 \newlabel{fig:tc_01_007}{{5.1}{3 5}{Průběh délky cyklu\relax \relax }{figure.caption.7}{}}203 \@writefile{lof}{\contentsline {figure}{\numberline {5.2}{\ignorespaces Pr\IeC {\r u}b\IeC {\v e}h sou\IeC {\v c}tu d\IeC {\'e}lek front\relax }}{3 5}{figure.caption.8}}204 \newlabel{fig:qsum_01_007}{{5.2}{3 5}{Průběh součtu délek front\relax \relax }{figure.caption.8}{}}205 \@writefile{lof}{\contentsline {figure}{\numberline {5.3}{\ignorespaces Po\IeC {\v c}et zastaven\IeC {\'\i }\relax }}{3 6}{figure.caption.9}}206 \newlabel{fig:tc_01_007}{{5.3}{3 6}{Počet zastavení\relax \relax }{figure.caption.9}{}}207 \@writefile{lof}{\contentsline {figure}{\numberline {5.4}{\ignorespaces Pr\IeC {\r u}m\IeC {\v e}rn\IeC {\'a} doba j\IeC {\'\i }zdy\relax }}{3 6}{figure.caption.10}}208 \newlabel{fig:qsum_01_007}{{5.4}{3 6}{Průměrná doba jízdy\relax \relax }{figure.caption.10}{}}209 \@writefile{toc}{\contentsline {section}{\numberline {5.2}Sc\IeC {\'e}n\IeC {\'a}\IeC {\v r} 2}{3 7}{section.5.2}}180 \@writefile{toc}{\contentsline {section}{\numberline {4.2}Simulace}{29}{section.4.2}} 181 \@writefile{toc}{\contentsline {subsubsection}{\numberline {4.2.0.2}VGS API}{30}{subsubsection.4.2.0.2}} 182 \newlabel{ss:vgs_api}{{4.2.0.2}{30}{VGS API\relax }{subsubsection.4.2.0.2}{}} 183 \@writefile{toc}{\contentsline {subsection}{\numberline {4.2.1}\IeC {\v R}adi\IeC {\v c}e}{31}{subsection.4.2.1}} 184 \@writefile{toc}{\contentsline {subsection}{\numberline {4.2.2}Oblast simulace}{31}{subsection.4.2.2}} 185 \newlabel{ss:oblast_simulace}{{4.2.2}{31}{Oblast simulace\relax }{subsection.4.2.2}{}} 186 \@writefile{lof}{\contentsline {figure}{\numberline {4.1}{\ignorespaces K\IeC {\v r}i\IeC {\v z}ovatka 601\relax }}{31}{figure.caption.5}} 187 \newlabel{fig:601}{{4.1}{31}{Křižovatka 601\relax \relax }{figure.caption.5}{}} 188 \citation{6_tuc_lq} 189 \@writefile{lof}{\contentsline {figure}{\numberline {4.2}{\ignorespaces K\IeC {\v r}i\IeC {\v z}ovatka 495\relax }}{32}{figure.caption.6}} 190 \newlabel{fig:601}{{4.2}{32}{Křižovatka 495\relax \relax }{figure.caption.6}{}} 191 \@writefile{toc}{\contentsline {section}{\numberline {4.3}Popis algoritmu}{32}{section.4.3}} 192 \@writefile{toc}{\contentsline {section}{\numberline {4.4}Mo\IeC {\v z}n\IeC {\'e} vylep\IeC {\v s}en\IeC {\'\i } do budoucna}{33}{section.4.4}} 193 \@writefile{toc}{\contentsline {subsubsection}{\numberline {4.4.0.1}Model toku}{33}{subsubsection.4.4.0.1}} 194 \newlabel{eq:teor_tok}{{4.27}{33}{Model toku\relax }{equation.4.4.27}{}} 195 \newlabel{eq:exp_tok}{{4.28}{33}{Model toku\relax }{equation.4.4.28}{}} 196 \newlabel{eq:lin_tok}{{4.31}{33}{Model toku\relax }{equation.4.4.31}{}} 197 \@writefile{toc}{\contentsline {subsection}{\numberline {4.4.1}Odhdad odbo\IeC {\v c}ovac\IeC {\'\i }ch pom\IeC {\v e}r\IeC {\r u}}{33}{subsection.4.4.1}} 198 \@writefile{toc}{\contentsline {chapter}{\numberline {5}V\IeC {\'y}sledky}{35}{chapter.5}} 199 \@writefile{lof}{\addvspace {10\p@ }} 200 \@writefile{lot}{\addvspace {10\p@ }} 201 \@writefile{toc}{\contentsline {section}{\numberline {5.1}Sc\IeC {\'e}n\IeC {\'a}\IeC {\v r} 1}{35}{section.5.1}} 202 \@writefile{lof}{\contentsline {figure}{\numberline {5.1}{\ignorespaces Pr\IeC {\r u}b\IeC {\v e}h d\IeC {\'e}lky cyklu\relax }}{36}{figure.caption.7}} 203 \newlabel{fig:tc_01_007}{{5.1}{36}{Průběh délky cyklu\relax \relax }{figure.caption.7}{}} 204 \@writefile{lof}{\contentsline {figure}{\numberline {5.2}{\ignorespaces Pr\IeC {\r u}b\IeC {\v e}h sou\IeC {\v c}tu d\IeC {\'e}lek front\relax }}{36}{figure.caption.8}} 205 \newlabel{fig:qsum_01_007}{{5.2}{36}{Průběh součtu délek front\relax \relax }{figure.caption.8}{}} 206 \@writefile{lof}{\contentsline {figure}{\numberline {5.3}{\ignorespaces Po\IeC {\v c}et zastaven\IeC {\'\i }\relax }}{37}{figure.caption.9}} 207 \newlabel{fig:tc_01_007}{{5.3}{37}{Počet zastavení\relax \relax }{figure.caption.9}{}} 208 \@writefile{lof}{\contentsline {figure}{\numberline {5.4}{\ignorespaces Pr\IeC {\r u}m\IeC {\v e}rn\IeC {\'a} doba j\IeC {\'\i }zdy\relax }}{37}{figure.caption.10}} 209 \newlabel{fig:qsum_01_007}{{5.4}{37}{Průměrná doba jízdy\relax \relax }{figure.caption.10}{}} 210 \@writefile{toc}{\contentsline {section}{\numberline {5.2}Sc\IeC {\'e}n\IeC {\'a}\IeC {\v r} 2}{38}{section.5.2}} 210 211 \bibstyle{czechiso} 211 212 \bibdata{Bibl/bibl} 212 \@writefile{toc}{\contentsline {chapter}{\numberline {6}Z\IeC {\'a}v\IeC {\v e}r}{3 8}{chapter.6}}213 \@writefile{toc}{\contentsline {chapter}{\numberline {6}Z\IeC {\'a}v\IeC {\v e}r}{39}{chapter.6}} 213 214 \@writefile{lof}{\addvspace {10\p@ }} 214 215 \@writefile{lot}{\addvspace {10\p@ }} … … 229 230 \bibcite{wooldridge}{15} 230 231 \citation{*} 231 \@writefile{toc}{\contentsline {chapter}{Literatura}{4 0}{chapter*.11}}232 \HyPL@Entry{4 7<</S/R>>}232 \@writefile{toc}{\contentsline {chapter}{Literatura}{41}{chapter*.11}} 233 \HyPL@Entry{48<</S/R>>} 233 234 \@writefile{toc}{\contentsline {chapter}{\numberline {A}P\IeC {\v r}\IeC {\'\i }loha 1}{I}{appendix.A}} 234 235 \@writefile{lof}{\addvspace {10\p@ }} -
applications/doprava/texty/novotny_vyzk_LQ/vyzk.log
r1425 r1427 1 This is pdfTeX, Version 3.1415926-1.40.10 (TeX Live 2009/Debian) (format=latex 2011.10.24) 3 FEB 2012 1 2:301 This is pdfTeX, Version 3.1415926-1.40.10 (TeX Live 2009/Debian) (format=latex 2011.10.24) 3 FEB 2012 14:39 2 2 entering extended mode 3 3 %&-line parsing enabled. … … 803 803 804 804 805 ]) (./ 02_Agents/Agents.tex [2]805 ]) (./MatematicMethods/MatematicMethods.tex [2] 806 806 Kapitola 2. 807 807 Package hyperref Info: bookmark level for unknown definition defaults to 0 on i … … 817 817 (Font) Font shape `OMS/cmsy/m/n' tried instead on input line 34. 818 818 [4] [5] 819 Kapitola 3. 820 [6 821 822 ] 823 Overfull \hbox (4.12453pt too wide) in paragraph at lines 137--140 819 Overfull \hbox (4.12453pt too wide) in paragraph at lines 134--137 824 820 []\OT1/cmr/m/n/12 Dv^^Te strate-gie, $\OML/cmm/m/it/12 a[]$ \OT1/cmr/m/n/12 a $ 825 821 \OML/cmm/m/it/12 a[]$ \OT1/cmr/m/n/12 jsou v Nashov^^Te rovnosti, pokud … … 836 832 [] 837 833 838 [ 7] [8]834 [6] [7] 839 835 Underfull \hbox (badness 10000) in paragraph at lines 74--80 840 836 841 837 [] 842 838 843 [ 9]839 [8] 844 840 845 841 Package amsmath Warning: Foreign command \over; … … 847 843 (amsmath) on input line 142. 848 844 849 ) (./Reinforcement_learning_pouziti/Reinforcement_learning_pouziti.tex [10] 850 Underfull \hbox (badness 10000) in paragraph at lines 28--31 851 852 [] 853 854 855 Overfull \hbox (22.51053pt too wide) in paragraph at lines 34--51 856 \OT1/cmr/m/n/12 kter^^Se jsou pro sim-u-laci po dlouhou dobu op-ti-mal-i-zov^^S 857 any a je-jich^^Tz nasazen^^S^^P zna^^Tcn^^Te zjednodu^^Tsuje 858 [] 859 860 [11]) (./03_RMM/RMM.tex [12] 845 [9]) (./03_RMM/RMM.tex 861 846 Overfull \hbox (31.25705pt too wide) in paragraph at lines 37--38 862 847 []\OT1/cmr/m/n/12 Rekursivn^^S^^P mod-e-lov^^Sa struk-tura $\OML/cmm/m/it/12 RM … … 864 849 [] 865 850 851 [10] 866 852 867 853 LaTeX Font Warning: Font shape `OT1/mnozina/m/n' in size <5> not available 868 854 (Font) size <6> substituted on input line 52. 869 855 870 [13] 856 871 857 Underfull \hbox (badness 10000) in paragraph at lines 72--76 872 858 … … 878 864 [] 879 865 880 866 [11] 881 867 Underfull \hbox (badness 10000) in paragraph at lines 82--84 882 868 … … 889 875 [] 890 876 891 [14]) (./04_Bayes/Bayes.tex877 ) (./04_Bayes/Bayes.tex 892 878 Overfull \hbox (2.46631pt too wide) in paragraph at lines 4--12 893 879 \OT1/cmr/m/n/12 V t^^Seto kapi-tole je nazna^^Tcena metoda bayesovsk^^Seho u^^T … … 895 881 [] 896 882 883 [12] 897 884 Package hyperref Info: bookmark level for unknown proposition defaults to 0 on 898 885 input line 19. 899 [15]) (./06_Bayes_rmm_pouziti/Bayes_rmm_pouziti.tex 900 Underfull \hbox (badness 10000) in paragraph at lines 5--9 901 902 [] 903 904 [16] 905 Overfull \hbox (6.66295pt too wide) in paragraph at lines 10--17 906 \OML/cmm/m/it/12 R[]; R[]$ \OT1/cmr/m/n/12 ovl^^Sadal jednu z nich. K mod-e-lov 907 ^^San^^S^^P chov^^San^^S^^P druh^^Seho agenta bylo pou^^Tzito t^^Tr^^S^^Psyup^^ 908 Tnov^^Se 909 [] 910 911 912 Underfull \hbox (badness 10000) in paragraph at lines 10--17 913 914 [] 915 916 File: ./06_Bayes_rmm_pouziti/rmm_structure.eps Graphic file (type eps) 917 <./06_Bayes_rmm_pouziti/rmm_structure.eps> 918 919 LaTeX Font Warning: Font shape `OT1/mnozina/m/n' in size <10.95> not available 920 (Font) size <10> substituted on input line 30. 921 922 [17] 923 File: ./06_Bayes_rmm_pouziti/rmm_results.eps Graphic file (type eps) 924 <./06_Bayes_rmm_pouziti/rmm_results.eps>) (./LQ_rizeni.tex 886 [13]) (./LQ_rizeni.tex 925 887 Overfull \hbox (9.84808pt too wide) in paragraph at lines 3--9 926 888 \OT1/cmr/m/n/12 obecn^^Te o metodu, kdy je syst^^Sem v diskr^^Setn^^S^^Pm ^^Tca … … 940 902 [] 941 903 942 [18] 943 Overfull \hbox (0.73724pt too wide) in paragraph at lines 37--42 904 [14]) (./Pouziti/Pouziti.tex [15] 905 Kapitola 3. 906 (./Reinforcement_learning_pouziti/Reinforcement_learning_pouziti.tex [16 907 908 ] 909 Underfull \hbox (badness 10000) in paragraph at lines 26--29 910 911 [] 912 913 ) (./06_Bayes_rmm_pouziti/Bayes_rmm_pouziti.tex 914 Underfull \hbox (badness 10000) in paragraph at lines 5--9 915 916 [] 917 918 919 Overfull \hbox (6.66295pt too wide) in paragraph at lines 10--17 920 \OML/cmm/m/it/12 R[]; R[]$ \OT1/cmr/m/n/12 ovl^^Sadal jednu z nich. K mod-e-lov 921 ^^San^^S^^P chov^^San^^S^^P druh^^Seho agenta bylo pou^^Tzito t^^Tr^^S^^Psyup^^ 922 Tnov^^Se 923 [] 924 925 926 Underfull \hbox (badness 10000) in paragraph at lines 10--17 927 928 [] 929 930 File: ./06_Bayes_rmm_pouziti/rmm_structure.eps Graphic file (type eps) 931 <./06_Bayes_rmm_pouziti/rmm_structure.eps> 932 933 LaTeX Font Warning: Font shape `OT1/mnozina/m/n' in size <10.95> not available 934 (Font) size <10> substituted on input line 30. 935 936 [17] [18] 937 File: ./06_Bayes_rmm_pouziti/rmm_results.eps Graphic file (type eps) 938 <./06_Bayes_rmm_pouziti/rmm_results.eps>) 939 Overfull \hbox (22.51053pt too wide) in paragraph at lines 12--29 940 \OT1/cmr/m/n/12 kter^^Se jsou pro sim-u-laci po dlouhou dobu op-ti-mal-i-zov^^S 941 any a je-jich^^Tz nasazen^^S^^P zna^^Tcn^^Te zjednodu^^Tsuje 942 [] 943 944 [19] 945 Overfull \hbox (0.73724pt too wide) in paragraph at lines 39--44 944 946 \OT1/cmr/m/n/12 LQ ^^Tr^^S^^Pzen^^S^^P bylo pou^^Tzito v [[]] k nalezen^^S^^P o 945 947 p-tim^^Saln^^S^^P d^^Selky ze-len^^Sych v syst^^Semu 13-ti sign^^Saln^^S^^Pch 946 948 [] 947 949 948 [ 19]949 Overfull \hbox (10.17679pt too wide) in paragraph at lines 97--101950 [20] 951 Overfull \hbox (10.17679pt too wide) in paragraph at lines 101--105 950 952 \OT1/cmr/m/n/12 zm^^Tenu d^^Selky trv^^San^^S^^P ze-len^^Se oproti nomin^^Saln^ 951 953 ^S^^Pm hod-not^^Sam. Kvadrat-ick^^Se krit^^Serium op-tim^^Saln^^S^^Pho 952 954 [] 953 955 954 [2 0]) (./Implementation/Implementation.tex [21]956 [21]) (./Implementation/Implementation.tex [22] 955 957 Kapitola 4. 956 958 … … 984 986 [] 985 987 986 [2 2988 [23 987 989 988 990 ] … … 990 992 991 993 992 [2 3] [24] (./Implementation/Minimalization.tex [25]993 994 LaTeX Warning: Reference `eq:prechod' on page 2 6undefined on input line 128.995 996 [2 6] [27]994 [24] [25] (./Implementation/Minimalization.tex [26] 995 996 LaTeX Warning: Reference `eq:prechod' on page 27 undefined on input line 128. 997 998 [27] [28] 997 999 Overfull \hbox (33.4813pt too wide) in paragraph at lines 191--192 998 1000 [][][][][][][][][][][][][][][][][][][][][][][][][] … … 1022 1024 [] 1023 1025 1024 [2 8]1026 [29] 1025 1027 Underfull \hbox (badness 10000) in paragraph at lines 99--123 1026 1028 … … 1032 1034 [] 1033 1035 1034 [ 29]1036 [30] 1035 1037 File: Implementation/fig/601.eps Graphic file (type eps) 1036 1038 <Implementation/fig/601.eps> 1037 1039 File: Implementation/fig/495.eps Graphic file (type eps) 1038 <Implementation/fig/495.eps> [3 0]1040 <Implementation/fig/495.eps> [31] 1039 1041 Underfull \hbox (badness 10000) in paragraph at lines 155--172 1040 1042 1041 1043 [] 1042 1044 1043 (./Implementation/ChangingFlow.tex [3 1]1045 (./Implementation/ChangingFlow.tex [32] 1044 1046 Underfull \hbox (badness 10000) in paragraph at lines 43--54 1045 1047 1046 1048 [] 1047 1049 1048 [3 2])) (./Results/Results.tex [33]1050 [33])) (./Results/Results.tex [34] 1049 1051 Kapitola 5. 1050 1052 File: Results/fig/tc_01_007.eps Graphic file (type eps) 1051 <Results/fig/tc_01_007.eps> [3 41053 <Results/fig/tc_01_007.eps> [35 1052 1054 1053 1055 ] 1054 1056 File: Results/fig/qsum_01_007.eps Graphic file (type eps) 1055 <Results/fig/qsum_01_007.eps> [3 5]1057 <Results/fig/qsum_01_007.eps> [36] 1056 1058 File: Results/fig/01/a_numStops.eps Graphic file (type eps) 1057 1059 … … 1059 1061 File: Results/fig/01/a_travelTimeAvg.eps Graphic file (type eps) 1060 1062 <Results/fig/01/a_travelTimeAvg.eps>) 1061 (./Ending.tex [3 6] [37]1063 (./Ending.tex [37] [38] 1062 1064 Kapitola 6. 1063 ) (./vyzk.bbl [3 81064 1065 ] [ 391066 1067 ]) [4 0] (./Appendix//01.tex1065 ) (./vyzk.bbl [39 1066 1067 ] [40 1068 1069 ]) [41] (./Appendix//01.tex 1068 1070 P\v {r}\'{\i }loha A. 1069 1071 ) … … 1087 1089 ) 1088 1090 Here is how much of TeX's memory you used: 1089 1061 4strings out of 4950621090 144 408string characters out of 11826431091 2315 94words of memory out of 30000001092 1353 4multiletter control sequences out of 15000+500001091 10619 strings out of 495062 1092 144524 string characters out of 1182643 1093 231586 words of memory out of 3000000 1094 13535 multiletter control sequences out of 15000+50000 1093 1095 17737 words of font info for 73 fonts, out of 3000000 for 9000 1094 1096 29 hyphenation exceptions out of 8191 1095 1097 43i,11n,45p,276b,1730s stack positions out of 5000i,500n,10000p,200000b,50000s 1096 1098 1097 Output written on vyzk.dvi (4 8 pages, 211484 bytes).1099 Output written on vyzk.dvi (49 pages, 213264 bytes). -
applications/doprava/texty/novotny_vyzk_LQ/vyzk.tex
r1425 r1427 21 21 \newcommand{\zamereni}{Tvorba softwaru} 22 22 % \def \Authors {Autor\?: Jakub Novotný Vedoucí práce\?: Ing. Václav Šmídl, Ph.D.} 23 \def \DatumDP {Praha, 201 1}23 \def \DatumDP {Praha, 2012} 24 24 \def \autor {Jakub Novotný} 25 25 %\def \vedouci {Ing. Václav Šmídl, Ph.D.} … … 119 119 %input fieles 120 120 \input{01_Intro/Intro.tex} 121 \input{ 02_Agents/Agents.tex}121 \input{MatematicMethods/MatematicMethods.tex} 122 122 \input{Reinforcement_learning.tex} 123 \input{Reinforcement_learning_pouziti/Reinforcement_learning_pouziti.tex} 123 124 124 \input{03_RMM/RMM.tex} 125 \input{04_Bayes/Bayes.tex} 126 \input{06_Bayes_rmm_pouziti/Bayes_rmm_pouziti.tex} 125 \input{04_Bayes/Bayes.tex} 126 \input{LQ_rizeni.tex} 127 \input{Pouziti/Pouziti.tex} 127 128 128 129 %\input{05_I_a_in_dec_t_c/I_a_in_dec_t_c.tex}130 131 132 \input{LQ_rizeni.tex}133 129 \input{Implementation/Implementation.tex} 134 130 \input{Results/Results.tex} -
applications/doprava/texty/novotny_vyzk_LQ/vyzk.thm
r1425 r1427 1 1 \contentsline {definition}{{Definice}{2.{1}}{Agent}}{3}{definition.2.1} 2 2 \contentsline {definition}{{Definice}{2.{2}}{Uspořádání na množině všech stavů}}{4}{definition.2.2} 3 \contentsline {definition}{{Definice}{ 3.{1}}{Dominance množiny}}{6}{definition.3.1}4 \contentsline {definition}{{Definice}{ 3.{2}}{Množina výsledků}}{7}{definition.3.2}5 \contentsline {definition}{{Definice}{ 3.{3}}{Dominance strategie}}{7}{definition.3.3}6 \contentsline {definition}{{Definice}{ 3.{4}}{Nashova rovnost}}{7}{definition.3.4}7 \contentsline {definition}{{Definice}{ 3.{5}}{Markovův rozhodovací proces}}{7}{definition.3.5}8 \contentsline {definition}{{Definice}{ 3.{6}}{V-funkce}}{8}{definition.3.6}9 \contentsline {definition}{{Definice}{ 3.{7}}{Q-function}}{8}{definition.3.7}10 \contentsline {definition}{{Definice}{ 3.{8}}{Bellmanova rovnice optimality}}{9}{definition.3.8}11 \contentsline {definition}{{Definice}{ 3.{9}}{MLM}}{10}{definition.3.9}12 \contentsline {definition}{{Definice}{ 3.{10}}{Matice zisků}}{12}{definition.3.10}13 \contentsline {definition}{{Definice}{ 3.{11}}{Rekursivní modelová struktura}}{13}{definition.3.11}14 \contentsline {definition}{{Definice}{ 3.{12}}{Rekursivní model}}{13}{definition.3.12}15 \contentsline {definition}{{Definice}{ 3.{13}}{Užitečnost}}{14}{definition.3.13}16 \contentsline {definition}{{Definice}{ 3.{14}}{Podmíněná pravděpodobnost}}{15}{definition.3.14}17 \contentsline {proposition}{{Věta}{ 3.{1}}{Bayesova věta}}{15}{proposition.3.1}18 \contentsline {definition}{{Definice}{ 3.{15}}{Věrohodnostní funkce}}{16}{definition.3.15}3 \contentsline {definition}{{Definice}{2.{3}}{Dominance množiny}}{5}{definition.2.3} 4 \contentsline {definition}{{Definice}{2.{4}}{Množina výsledků}}{6}{definition.2.4} 5 \contentsline {definition}{{Definice}{2.{5}}{Dominance strategie}}{6}{definition.2.5} 6 \contentsline {definition}{{Definice}{2.{6}}{Nashova rovnost}}{6}{definition.2.6} 7 \contentsline {definition}{{Definice}{2.{7}}{Markovův rozhodovací proces}}{7}{definition.2.7} 8 \contentsline {definition}{{Definice}{2.{8}}{V-funkce}}{7}{definition.2.8} 9 \contentsline {definition}{{Definice}{2.{9}}{Q-function}}{8}{definition.2.9} 10 \contentsline {definition}{{Definice}{2.{10}}{Bellmanova rovnice optimality}}{8}{definition.2.10} 11 \contentsline {definition}{{Definice}{2.{11}}{MLM}}{9}{definition.2.11} 12 \contentsline {definition}{{Definice}{2.{12}}{Matice zisků}}{10}{definition.2.12} 13 \contentsline {definition}{{Definice}{2.{13}}{Rekursivní modelová struktura}}{11}{definition.2.13} 14 \contentsline {definition}{{Definice}{2.{14}}{Rekursivní model}}{11}{definition.2.14} 15 \contentsline {definition}{{Definice}{2.{15}}{Užitečnost}}{12}{definition.2.15} 16 \contentsline {definition}{{Definice}{2.{16}}{Podmíněná pravděpodobnost}}{13}{definition.2.16} 17 \contentsline {proposition}{{Věta}{2.{1}}{Bayesova věta}}{13}{proposition.2.1} 18 \contentsline {definition}{{Definice}{2.{17}}{Věrohodnostní funkce}}{14}{definition.2.17} -
applications/doprava/texty/novotny_vyzk_LQ/vyzk.toc
r1425 r1427 1 1 \select@language {czech} 2 2 \contentsline {chapter}{\numberline {1}\IeC {\'U}vod}{1}{chapter.1} 3 \contentsline {chapter}{\numberline {2}M ultiagentn\IeC {\'\i } syst\IeC {\'e}my}{3}{chapter.2}4 \contentsline {section}{\numberline {2.1} \IeC {\'U}vod}{3}{section.2.1}3 \contentsline {chapter}{\numberline {2}Matematick\IeC {\'e} metody rozhodov\IeC {\'a}n\IeC {\'\i }}{3}{chapter.2} 4 \contentsline {section}{\numberline {2.1}Multiagentn\IeC {\'\i } syst\IeC {\'e}my}{3}{section.2.1} 5 5 \contentsline {subsection}{\numberline {2.1.1}Historie}{3}{subsection.2.1.1} 6 6 \contentsline {subsection}{\numberline {2.1.2}Agent}{3}{subsection.2.1.2} 7 \contentsline {s ection}{\numberline {2.2}Druhy prost\IeC {\v r}ed\IeC {\'\i }}{4}{section.2.2}8 \contentsline {s ection}{\numberline {2.3}Interakce agent\IeC {\r u}}{4}{section.2.3}9 \contentsline {s ubsection}{\numberline {2.3.1}Stavy prost\IeC {\v r}ed\IeC {\'\i } a preference agent\IeC {\r u}}{4}{subsection.2.3.1}10 \contentsline { chapter}{\numberline {3}V\IeC {\'y}b\IeC {\v e}r strategie genta}{6}{chapter.3}11 \contentsline {s ection}{\numberline {3.1}V\IeC {\'y}b\IeC {\v e}r strategie podle teorie her}{6}{section.3.1}12 \contentsline {s ection}{\numberline {3.2}Markov\IeC {\r u}v rozhodvac\IeC {\'\i } proces}{7}{section.3.2}13 \contentsline {subs ubsection}{\numberline {3.2.0.1}Dynamick\IeC {\'e} programov\IeC {\'a}n\IeC {\'\i }}{8}{subsubsection.3.2.0.1}14 \contentsline {subs ection}{\numberline {3.2.1}Zp\IeC {\v e}tnovazebn\IeC {\'e} u\IeC {\v c}en\IeC {\'\i } (Reinforcement learning)}{9}{subsection.3.2.1}15 \contentsline {subsubsection}{\numberline { 3.2.1.1}Q-u\IeC {\v c}en\IeC {\'\i } (Q-learning)}{10}{subsubsection.3.2.1.1}16 \contentsline {s ubsection}{\numberline {3.2.2}U\IeC {\v c}en\IeC {\'\i } na z\IeC {\'a}klad\IeC {\v e} modelu (Model-based learning)}{10}{subsection.3.2.2}17 \contentsline {subsection}{\numberline { 3.2.3}Pou\IeC {\v z}it\IeC {\'\i } zp\IeC {\v e}tnovazebn\IeC {\'e}ho u\IeC {\v c}en\IeC {\'\i }}{11}{subsection.3.2.3}18 \contentsline {subs ubsection}{\numberline {3.2.3.1}Zp\IeC {\v e}tnovazebn\IeC {\'e} u\IeC {\v c}en\IeC {\'\i } na z\IeC {\'a}klad\IeC {\v e} modelu}{11}{subsubsection.3.2.3.1}19 \contentsline {s ubsection}{\numberline {3.2.4}Zhodnocen\IeC {\'\i }}{11}{subsection.3.2.4}20 \contentsline {s ection}{\numberline {3.3}RMM - Rekurzivn\IeC {\'\i } modelov\IeC {\'e} metody}{12}{section.3.3}21 \contentsline {s ubsection}{\numberline {3.3.1}Form\IeC {\'a}ln\IeC {\'\i } definice}{12}{subsection.3.3.1}22 \contentsline { subsection}{\numberline {3.3.2}Rozhodovac\IeC {\'\i } algoritmus}{14}{subsection.3.3.2}23 \contentsline {s ubsection}{\numberline {3.3.3}Bayesovsk\IeC {\'e} u\IeC {\v c}en\IeC {\'\i }}{15}{subsection.3.3.3}24 \contentsline {s ubsubsection}{\numberline {3.3.3.1}V\IeC {\v e}rohodnostn\IeC {\'\i } funkce}{16}{subsubsection.3.3.3.1}25 \contentsline {s ubsection}{\numberline {3.3.4}Pou\IeC {\v z}it\IeC {\'\i } RMM a Bayesova u\IeC {\v c}en\IeC {\'\i } v decentralizovan\IeC {\'e}m \IeC {\v r}\IeC {\'\i }zen\IeC {\'\i } dopravy}{16}{subsection.3.3.4}26 \contentsline {subsection}{\numberline {3.3. 5}Zhodnocen\IeC {\'\i }}{18}{subsection.3.3.5}27 \contentsline {s ection}{\numberline {3.4}LQ \IeC {\v r}\IeC {\'\i }zen\IeC {\'\i }}{18}{section.3.4}28 \contentsline {subsection}{\numberline {3. 4.1}Pou\IeC {\v z}it\IeC {\'\i } LQ \IeC {\v r}\IeC {\'\i }zen\IeC {\'\i } ve strategii TUC}{19}{subsection.3.4.1}29 \contentsline {subsubsection}{\numberline {3. 4.1.1}Kvadratick\IeC {\'e} krit\IeC {\'e}rium}{21}{subsubsection.3.4.1.1}30 \contentsline {chapter}{\numberline {4}Implementace}{2 2}{chapter.4}31 \contentsline {section}{\numberline {4.1}Pou\IeC {\v z}it\IeC {\'a} metoda}{2 4}{section.4.1}32 \contentsline {subsection}{\numberline {4.1.1}P\IeC {\v r}echodov\IeC {\'e} vztahy}{2 4}{subsection.4.1.1}33 \contentsline {subsection}{\numberline {4.1.2}Minimalizace krit\IeC {\'e}ria}{2 5}{subsection.4.1.2}34 \contentsline {subsubsection}{\numberline {4.1.2.1}Implementace minimalizace}{2 7}{subsubsection.4.1.2.1}35 \contentsline {section}{\numberline {4.2}Simulace}{2 8}{section.4.2}36 \contentsline {subsubsection}{\numberline {4.2.0.2}VGS API}{ 29}{subsubsection.4.2.0.2}37 \contentsline {subsection}{\numberline {4.2.1}\IeC {\v R}adi\IeC {\v c}e}{3 0}{subsection.4.2.1}38 \contentsline {subsection}{\numberline {4.2.2}Oblast simulace}{3 0}{subsection.4.2.2}39 \contentsline {section}{\numberline {4.3}Popis algoritmu}{3 1}{section.4.3}40 \contentsline {section}{\numberline {4.4}Mo\IeC {\v z}n\IeC {\'e} vylep\IeC {\v s}en\IeC {\'\i } do budoucna}{3 2}{section.4.4}41 \contentsline {subsubsection}{\numberline {4.4.0.1}Model toku}{3 2}{subsubsection.4.4.0.1}42 \contentsline {subsection}{\numberline {4.4.1}Odhdad odbo\IeC {\v c}ovac\IeC {\'\i }ch pom\IeC {\v e}r\IeC {\r u}}{3 2}{subsection.4.4.1}43 \contentsline {chapter}{\numberline {5}V\IeC {\'y}sledky}{3 4}{chapter.5}44 \contentsline {section}{\numberline {5.1}Sc\IeC {\'e}n\IeC {\'a}\IeC {\v r} 1}{3 4}{section.5.1}45 \contentsline {section}{\numberline {5.2}Sc\IeC {\'e}n\IeC {\'a}\IeC {\v r} 2}{3 7}{section.5.2}46 \contentsline {chapter}{\numberline {6}Z\IeC {\'a}v\IeC {\v e}r}{3 8}{chapter.6}47 \contentsline {chapter}{Literatura}{4 0}{chapter*.11}7 \contentsline {subsection}{\numberline {2.1.3}Druhy prost\IeC {\v r}ed\IeC {\'\i }}{4}{subsection.2.1.3} 8 \contentsline {subsection}{\numberline {2.1.4}Stavy prost\IeC {\v r}ed\IeC {\'\i } a preference agent\IeC {\r u}}{4}{subsection.2.1.4} 9 \contentsline {section}{\numberline {2.2}V\IeC {\'y}b\IeC {\v e}r strategie podle teorie her}{5}{section.2.2} 10 \contentsline {section}{\numberline {2.3}Zp\IeC {\v e}tnovazebn\IeC {\'e} u\IeC {\v c}en\IeC {\'\i }}{6}{section.2.3} 11 \contentsline {subsection}{\numberline {2.3.1}Markov\IeC {\r u}v rozhodvac\IeC {\'\i } proces}{6}{subsection.2.3.1} 12 \contentsline {subsection}{\numberline {2.3.2}Dynamick\IeC {\'e} programov\IeC {\'a}n\IeC {\'\i }}{7}{subsection.2.3.2} 13 \contentsline {subsection}{\numberline {2.3.3}Zp\IeC {\v e}tnovazebn\IeC {\'e} u\IeC {\v c}en\IeC {\'\i } (Reinforcement learning)}{9}{subsection.2.3.3} 14 \contentsline {subsubsection}{\numberline {2.3.3.1}Q-u\IeC {\v c}en\IeC {\'\i } (Q-learning)}{9}{subsubsection.2.3.3.1} 15 \contentsline {subsubsection}{\numberline {2.3.3.2}U\IeC {\v c}en\IeC {\'\i } na z\IeC {\'a}klad\IeC {\v e} modelu (Model-based learning)}{9}{subsubsection.2.3.3.2} 16 \contentsline {section}{\numberline {2.4}RMM - Rekurzivn\IeC {\'\i } modelov\IeC {\'e} metody}{10}{section.2.4} 17 \contentsline {subsection}{\numberline {2.4.1}Form\IeC {\'a}ln\IeC {\'\i } definice}{10}{subsection.2.4.1} 18 \contentsline {subsection}{\numberline {2.4.2}Rozhodovac\IeC {\'\i } algoritmus}{12}{subsection.2.4.2} 19 \contentsline {section}{\numberline {2.5}Bayesovsk\IeC {\'e} u\IeC {\v c}en\IeC {\'\i }}{13}{section.2.5} 20 \contentsline {subsubsection}{\numberline {2.5.0.1}V\IeC {\v e}rohodnostn\IeC {\'\i } funkce}{13}{subsubsection.2.5.0.1} 21 \contentsline {section}{\numberline {2.6}LQ \IeC {\v r}\IeC {\'\i }zen\IeC {\'\i }}{14}{section.2.6} 22 \contentsline {chapter}{\numberline {3}Pou\IeC {\v z}it\IeC {\'\i } rozhodovac\IeC {\'\i }ch metod v \IeC {\v r}\IeC {\'\i }zen\IeC {\'\i } dopravy}{16}{chapter.3} 23 \contentsline {section}{\numberline {3.1}Pou\IeC {\v z}it\IeC {\'\i } zp\IeC {\v e}tnovazebn\IeC {\'e}ho u\IeC {\v c}en\IeC {\'\i }}{16}{section.3.1} 24 \contentsline {section}{\numberline {3.2}Pou\IeC {\v z}it\IeC {\'\i } RMM a Bayesova u\IeC {\v c}en\IeC {\'\i } v decentralizovan\IeC {\'e}m \IeC {\v r}\IeC {\'\i }zen\IeC {\'\i } dopravy}{17}{section.3.2} 25 \contentsline {section}{\numberline {3.3}Zhodnocen\IeC {\'\i }}{19}{section.3.3} 26 \contentsline {subsection}{\numberline {3.3.1}Pou\IeC {\v z}it\IeC {\'\i } zp\IeC {\v e}tnovazebn\IeC {\'e}ho u\IeC {\v c}en\IeC {\'\i }}{19}{subsection.3.3.1} 27 \contentsline {subsection}{\numberline {3.3.2}Pou\IeC {\v z}it\IeC {\'\i } RMM a Bayesova u\IeC {\v c}en\IeC {\'\i }}{20}{subsection.3.3.2} 28 \contentsline {subsection}{\numberline {3.3.3}Pou\IeC {\v z}it\IeC {\'\i } LQ \IeC {\v r}\IeC {\'\i }zen\IeC {\'\i } ve strategii TUC}{20}{subsection.3.3.3} 29 \contentsline {subsubsection}{\numberline {3.3.3.1}Kvadratick\IeC {\'e} krit\IeC {\'e}rium}{21}{subsubsection.3.3.3.1} 30 \contentsline {chapter}{\numberline {4}Implementace}{23}{chapter.4} 31 \contentsline {section}{\numberline {4.1}Pou\IeC {\v z}it\IeC {\'a} metoda}{25}{section.4.1} 32 \contentsline {subsection}{\numberline {4.1.1}P\IeC {\v r}echodov\IeC {\'e} vztahy}{25}{subsection.4.1.1} 33 \contentsline {subsection}{\numberline {4.1.2}Minimalizace krit\IeC {\'e}ria}{26}{subsection.4.1.2} 34 \contentsline {subsubsection}{\numberline {4.1.2.1}Implementace minimalizace}{28}{subsubsection.4.1.2.1} 35 \contentsline {section}{\numberline {4.2}Simulace}{29}{section.4.2} 36 \contentsline {subsubsection}{\numberline {4.2.0.2}VGS API}{30}{subsubsection.4.2.0.2} 37 \contentsline {subsection}{\numberline {4.2.1}\IeC {\v R}adi\IeC {\v c}e}{31}{subsection.4.2.1} 38 \contentsline {subsection}{\numberline {4.2.2}Oblast simulace}{31}{subsection.4.2.2} 39 \contentsline {section}{\numberline {4.3}Popis algoritmu}{32}{section.4.3} 40 \contentsline {section}{\numberline {4.4}Mo\IeC {\v z}n\IeC {\'e} vylep\IeC {\v s}en\IeC {\'\i } do budoucna}{33}{section.4.4} 41 \contentsline {subsubsection}{\numberline {4.4.0.1}Model toku}{33}{subsubsection.4.4.0.1} 42 \contentsline {subsection}{\numberline {4.4.1}Odhdad odbo\IeC {\v c}ovac\IeC {\'\i }ch pom\IeC {\v e}r\IeC {\r u}}{33}{subsection.4.4.1} 43 \contentsline {chapter}{\numberline {5}V\IeC {\'y}sledky}{35}{chapter.5} 44 \contentsline {section}{\numberline {5.1}Sc\IeC {\'e}n\IeC {\'a}\IeC {\v r} 1}{35}{section.5.1} 45 \contentsline {section}{\numberline {5.2}Sc\IeC {\'e}n\IeC {\'a}\IeC {\v r} 2}{38}{section.5.2} 46 \contentsline {chapter}{\numberline {6}Z\IeC {\'a}v\IeC {\v e}r}{39}{chapter.6} 47 \contentsline {chapter}{Literatura}{41}{chapter*.11} 48 48 \contentsline {chapter}{\numberline {A}P\IeC {\v r}\IeC {\'\i }loha 1}{I}{appendix.A}