39cd826ea3
git-svn-id: svn://scm.gforge.inria.fr/svnroot/paradiseo@661 331e1502-861f-0410-8da2-ba01fb791d7f
145 lines
9.2 KiB
TeX
145 lines
9.2 KiB
TeX
\hypertarget{classpeoEA}{
|
|
\section{peo\-EA$<$ EOT $>$ Class Template Reference}
|
|
\label{classpeoEA}\index{peoEA@{peoEA}}
|
|
}
|
|
The \hyperlink{classpeoEA}{peo\-EA} class offers an elementary evolutionary algorithm implementation.
|
|
|
|
|
|
{\tt \#include $<$peo\-EA.h$>$}
|
|
|
|
Inheritance diagram for peo\-EA$<$ EOT $>$::\begin{figure}[H]
|
|
\begin{center}
|
|
\leavevmode
|
|
\includegraphics[height=3cm]{classpeoEA}
|
|
\end{center}
|
|
\end{figure}
|
|
\subsection*{Public Member Functions}
|
|
\begin{CompactItemize}
|
|
\item
|
|
\hyperlink{classpeoEA_dbfc4f8907bef234602149229f132371}{peo\-EA} (\bf{eo\-Continue}$<$ EOT $>$ \&\_\-\_\-cont, \hyperlink{classpeoPopEval}{peo\-Pop\-Eval}$<$ EOT $>$ \&\_\-\_\-pop\_\-eval, \bf{eo\-Select}$<$ EOT $>$ \&\_\-\_\-select, \hyperlink{classpeoTransform}{peo\-Transform}$<$ EOT $>$ \&\_\-\_\-trans, \bf{eo\-Replacement}$<$ EOT $>$ \&\_\-\_\-replace)
|
|
\begin{CompactList}\small\item\em Constructor for the evolutionary algorithm object - several basic parameters have to be specified, allowing for different levels of parallelism. \item\end{CompactList}\item
|
|
\hypertarget{classpeoEA_6ab8c321d29350634143a2a01cf2ad24}{
|
|
void \hyperlink{classpeoEA_6ab8c321d29350634143a2a01cf2ad24}{run} ()}
|
|
\label{classpeoEA_6ab8c321d29350634143a2a01cf2ad24}
|
|
|
|
\begin{CompactList}\small\item\em Evolutionary algorithm function - a side effect of the fact that the class is derived from the {\bf \hyperlink{classRunner}{Runner}} class, thus requiring the existence of a {\em run\/} function, the algorithm being executed on a distinct thread. \item\end{CompactList}\item
|
|
void \hyperlink{classpeoEA_3c709e3b2491147d26fee36138644613}{operator()} (\bf{eo\-Pop}$<$ EOT $>$ \&\_\-\_\-pop)
|
|
\begin{CompactList}\small\item\em \doxyref{Function} operator for specifying the population to be associated with the algorithm. \item\end{CompactList}\end{CompactItemize}
|
|
\subsection*{Private Attributes}
|
|
\begin{CompactItemize}
|
|
\item
|
|
\hypertarget{classpeoEA_5f015eebf42f176b9fe322488c446c2a}{
|
|
\bf{eo\-Continue}$<$ EOT $>$ \& \hyperlink{classpeoEA_5f015eebf42f176b9fe322488c446c2a}{cont}}
|
|
\label{classpeoEA_5f015eebf42f176b9fe322488c446c2a}
|
|
|
|
\item
|
|
\hypertarget{classpeoEA_9140259f50c9186edcb062b023624c96}{
|
|
\hyperlink{classpeoPopEval}{peo\-Pop\-Eval}$<$ EOT $>$ \& \hyperlink{classpeoEA_9140259f50c9186edcb062b023624c96}{pop\_\-eval}}
|
|
\label{classpeoEA_9140259f50c9186edcb062b023624c96}
|
|
|
|
\item
|
|
\hypertarget{classpeoEA_2d8428d69fdd6aefefbaf543fdd46d19}{
|
|
\bf{eo\-Select}$<$ EOT $>$ \& \hyperlink{classpeoEA_2d8428d69fdd6aefefbaf543fdd46d19}{select}}
|
|
\label{classpeoEA_2d8428d69fdd6aefefbaf543fdd46d19}
|
|
|
|
\item
|
|
\hypertarget{classpeoEA_713c77935eb8aafebfb9488cfaa4a363}{
|
|
\hyperlink{classpeoTransform}{peo\-Transform}$<$ EOT $>$ \& \hyperlink{classpeoEA_713c77935eb8aafebfb9488cfaa4a363}{trans}}
|
|
\label{classpeoEA_713c77935eb8aafebfb9488cfaa4a363}
|
|
|
|
\item
|
|
\hypertarget{classpeoEA_9bd2d4356cf7e69e3141dc269213aa8a}{
|
|
\bf{eo\-Replacement}$<$ EOT $>$ \& \hyperlink{classpeoEA_9bd2d4356cf7e69e3141dc269213aa8a}{replace}}
|
|
\label{classpeoEA_9bd2d4356cf7e69e3141dc269213aa8a}
|
|
|
|
\item
|
|
\hypertarget{classpeoEA_c0b110e410bc16283e8339f24b733772}{
|
|
\bf{eo\-Pop}$<$ EOT $>$ $\ast$ \hyperlink{classpeoEA_c0b110e410bc16283e8339f24b733772}{pop}}
|
|
\label{classpeoEA_c0b110e410bc16283e8339f24b733772}
|
|
|
|
\end{CompactItemize}
|
|
|
|
|
|
\subsection{Detailed Description}
|
|
\subsubsection*{template$<$class EOT$>$ class peo\-EA$<$ EOT $>$}
|
|
|
|
The \hyperlink{classpeoEA}{peo\-EA} class offers an elementary evolutionary algorithm implementation.
|
|
|
|
In addition, as compared with the algorithms provided by the \doxyref{EO} framework, the \hyperlink{classpeoEA}{peo\-EA} class has the underlying necessary structure for including, for example, parallel evaluation and parallel transformation operators, migration operators etc. Although there is no restriction on using the algorithms provided by the \doxyref{EO} framework, the drawback resides in the fact that the \doxyref{EO} implementation is exclusively sequential and, in consequence, no parallelism is provided. A simple example for constructing a \hyperlink{classpeoEA}{peo\-EA} object:
|
|
|
|
\begin{TabularC}{2}
|
|
\hline
|
|
... ~ &~ \\\hline
|
|
eo\-Pop$<$ EOT $>$ population( POP\_\-SIZE, pop\-Initializer ); ~ &// creation of a population with POP\_\-SIZE individuals - the pop\-Initializer is a functor to be called for each individual \\\hline
|
|
~ &~ \\\hline
|
|
eo\-Gen\-Continue$<$ EOT $>$ ea\-Cont( NUM\_\-GEN ); ~ &// number of generations for the evolutionary algorithm \\\hline
|
|
eo\-Check\-Point$<$ EOT $>$ ea\-Checkpoint\-Continue( ea\-Cont ); ~ &// checkpoint incorporating the continuation criterion - startpoint for adding other checkpoint objects \\\hline
|
|
~ &~ \\\hline
|
|
peo\-Seq\-Pop\-Eval$<$ EOT $>$ ea\-Pop\-Eval( eval\-Function ); ~ &// sequential evaluation functor wrapper - eval\-Function represents the actual evaluation functor \\\hline
|
|
~ &~ \\\hline
|
|
eo\-Ranking\-Select$<$ EOT $>$ selection\-Strategy; ~ &// selection strategy for creating the offspring population - a simple ranking selection in this case \\\hline
|
|
eo\-Select\-Number$<$ EOT $>$ ea\-Select( selection\-Strategy, POP\_\-SIZE ); ~ &// the number of individuals to be selected for creating the offspring population \\\hline
|
|
eo\-Ranking\-Select$<$ EOT $>$ selection\-Strategy; ~ &// selection strategy for creating the offspring population - a simple ranking selection in this case \\\hline
|
|
~ &~ \\\hline
|
|
eo\-SGATransform$<$ EOT $>$ transform( crossover, CROSS\_\-RATE, mutation, MUT\_\-RATE ); ~ &// transformation operator - crossover and mutation operators with their associated probabilities \\\hline
|
|
peo\-Seq\-Transform$<$ EOT $>$ ea\-Transform( transform ); ~ &// Paradis\-EO specific sequential operator - a parallel version may be specified in the same manner \\\hline
|
|
~ &~ \\\hline
|
|
eo\-Plus\-Replacement$<$ EOT $>$ ea\-Replace; ~ &// replacement strategy - for integrating the offspring resulting individuals in the initial population \\\hline
|
|
~ &~ \\\hline
|
|
peo\-EA$<$ EOT $>$ ea\-Alg( ea\-Checkpoint\-Continue, ea\-Pop\-Eval, ea\-Select, ea\-Transform, ea\-Replace ); ~ &// Paradis\-EO evolutionary algorithm integrating the above defined objects \\\hline
|
|
ea\-Alg( population ); ~ &// specifying the initial population for the algorithm \\\hline
|
|
... ~ &~ \\\hline
|
|
\end{TabularC}
|
|
|
|
|
|
|
|
|
|
Definition at line 82 of file peo\-EA.h.
|
|
|
|
\subsection{Constructor \& Destructor Documentation}
|
|
\hypertarget{classpeoEA_dbfc4f8907bef234602149229f132371}{
|
|
\index{peoEA@{peo\-EA}!peoEA@{peoEA}}
|
|
\index{peoEA@{peoEA}!peoEA@{peo\-EA}}
|
|
\subsubsection[peoEA]{\setlength{\rightskip}{0pt plus 5cm}template$<$class EOT$>$ \hyperlink{classpeoEA}{peo\-EA}$<$ EOT $>$::\hyperlink{classpeoEA}{peo\-EA} (\bf{eo\-Continue}$<$ EOT $>$ \& {\em \_\-\_\-cont}, \hyperlink{classpeoPopEval}{peo\-Pop\-Eval}$<$ EOT $>$ \& {\em \_\-\_\-pop\_\-eval}, \bf{eo\-Select}$<$ EOT $>$ \& {\em \_\-\_\-select}, \hyperlink{classpeoTransform}{peo\-Transform}$<$ EOT $>$ \& {\em \_\-\_\-trans}, \bf{eo\-Replacement}$<$ EOT $>$ \& {\em \_\-\_\-replace})}}
|
|
\label{classpeoEA_dbfc4f8907bef234602149229f132371}
|
|
|
|
|
|
Constructor for the evolutionary algorithm object - several basic parameters have to be specified, allowing for different levels of parallelism.
|
|
|
|
Depending on the requirements, a sequential or a parallel evaluation operator may be specified or, in the same manner, a sequential or a parallel transformation operator may be given as parameter. Out of the box objects may be provided, from the \doxyref{EO} package, for example, or custom defined ones may be specified, provided that they are derived from the correct base classes.
|
|
|
|
\begin{Desc}
|
|
\item[Parameters:]
|
|
\begin{description}
|
|
\item[{\em eo\-Continue$<$}]EOT $>$\& \_\-\_\-cont - continuation criterion specifying whether the algorithm should continue or not; \item[{\em peo\-Pop\-Eval$<$}]EOT $>$\& \_\-\_\-pop\_\-eval - evaluation operator; it allows the specification of parallel evaluation operators, aggregate evaluation functions, etc.; \item[{\em eo\-Select$<$}]EOT $>$\& \_\-\_\-select - selection strategy to be applied for constructing a list of offspring individuals; \item[{\em peo\-Transform$<$}]EOT $>$\& \_\-\_\-trans - transformation operator, i.e. crossover and mutation; allows for sequential or parallel transform; \item[{\em eo\-Replacement$<$}]EOT $>$\& \_\-\_\-replace - replacement strategy for integrating the offspring individuals in the initial population; \end{description}
|
|
\end{Desc}
|
|
|
|
|
|
Definition at line 126 of file peo\-EA.h.
|
|
|
|
References peo\-EA$<$ EOT $>$::pop\_\-eval, and peo\-EA$<$ EOT $>$::trans.
|
|
|
|
\subsection{Member Function Documentation}
|
|
\hypertarget{classpeoEA_3c709e3b2491147d26fee36138644613}{
|
|
\index{peoEA@{peo\-EA}!operator()@{operator()}}
|
|
\index{operator()@{operator()}!peoEA@{peo\-EA}}
|
|
\subsubsection[operator()]{\setlength{\rightskip}{0pt plus 5cm}template$<$class EOT$>$ void \hyperlink{classpeoEA}{peo\-EA}$<$ EOT $>$::operator() (\bf{eo\-Pop}$<$ EOT $>$ \& {\em \_\-\_\-pop})}}
|
|
\label{classpeoEA_3c709e3b2491147d26fee36138644613}
|
|
|
|
|
|
\doxyref{Function} operator for specifying the population to be associated with the algorithm.
|
|
|
|
\begin{Desc}
|
|
\item[Parameters:]
|
|
\begin{description}
|
|
\item[{\em eo\-Pop$<$}]EOT $>$\& \_\-\_\-pop - initial population of the algorithm, to be iteratively evolved; \end{description}
|
|
\end{Desc}
|
|
|
|
|
|
Definition at line 142 of file peo\-EA.h.
|
|
|
|
References peo\-EA$<$ EOT $>$::pop.
|
|
|
|
The documentation for this class was generated from the following file:\begin{CompactItemize}
|
|
\item
|
|
peo\-EA.h\end{CompactItemize}
|