\draw[blue,fill=blue!20,opacity=0.8] (O) -- (K) -- (M) -- (I) -- cycle;
\draw[blue,fill=blue!20,opacity=0.8] (K) -- (M) -- (N) -- (L) -- cycle;
\draw[blue,fill=blue!20,opacity=0.8] (I) -- (M) -- (N) -- (J) -- cycle;
-
+ \draw[->,thick,red] (O) -- (N);
%% Following is for debugging purposes so you can see where the points are
%% These are last so that they show up on top
%\foreach \xy in {O, A, B, C, D, E, F, G, H, I, J, K, L, M, N}{
%------------------------------------------------
\begin{frame}
-\frametitle{EP Heuristic, Restrictions and Variants}
-TODO Bells and whistles - all options. The description of each option goes to
-the appendix on slide \ref{appendix}.
+\frametitle{EP Heuristic, Restrictions}
+Input: an order containing \textbf{N} items. Output: a list of \textbf{M} boxes
+packed with the input items.
+\begin{columns}[T]
+ \begin{column}{0.5\textwidth}
+ Precondition: all items are sorted in decreasing order.
+ Criteria:
+ \begin{itemize}
+ \item area
+ \item height
+ \item first height, then volume
+ \item first volume, then height
+ \item first area, then height
+ \item first height, then area
+ \end{itemize}
+
+ \end{column}
+ \begin{column}{0.5\textwidth}
+ Postconditions:
+ \begin{itemize}
+ \item successor items are of sizes less or equal to predecessor ones
+ \item all items keep their orientation after insertion into a box
+ \end{itemize}
+ \end{column}
+\end{columns}
+
\end{frame}
+%------------------------------------------------
+
+\begin{frame}
+\frametitle{EP Heuristic, Variants}
+Input: an order containing \textbf{N} items. Output: a list of \textbf{M} boxes
+packed with the input items.
+\begin{columns}[T]
+ \begin{column}{0.5\textwidth}
+ EP-FFD: first fit.
+ \begin{itemize}
+ \item Put an item into a box
+ \item If the current box is too full for the item, pick a new box
+ \item Time complexity: $\Omega(3\cdot M \cdot N) \subseteq O(N^2)$
+ \end{itemize}
+
+ \end{column}
+ \begin{column}{0.5\textwidth}
+ EP-BFD: best fit. Before inserting the current item,
+ \begin{itemize}
+ \item if no previously packed box has enough room, place the item
+ in a new box
+ \item otherwise, pick one which maximizes $f_m$
+ \item Time complexity: depends on $f_m$
+ \end{itemize}
+ \end{column}
+\end{columns}
+The merit function $f_m$ is one of: maximize free volume after item insertion,
+minimize packing size, minimize packing size leveled, maximize residual spaces
+- details on slide \ref{appendix}.
+
+\end{frame}
%------------------------------------------------
\begin{itemize}
\item Implemented using Python3, standard libraries for (algebraic)
computation, data structures, sorting
- \item Disadvantage: runtime overhead, no third-party optimizer support
+ \item Disadvantages: runtime overhead, no third-party optimizer support
\item Advantages: portable, 100\%
\href{https://opensource.com/article/17/8/enterprise-open-source-advantages}{
FOSS} \footnote{\url{https://opensource.org/faq}}
\section{Appendix}
+\begin{frame}[allowframebreaks]
+\frametitle{EP heuristic, merit functions}
+\label{appendix}
+Convention: the parts after $max$ are introduced by the complexity of each
+merit function.
+\begin{enumerate}[1.]
+ \item Maximize free volume: pick the box which would be left with the most
+ free volume after accommodating the item. Time complexity: $O(N^2 +
+ max\{N,
+ N\})$ = $O(N^2)$
+ \item Minimize the maximum packing size: choose the box where either the
+ item is placed on top or, if not possible, the box with the most free
+ surface. Time complexity: as above
+ \item Level the EPs: choose the box whose EPs will have the least increase
+ in height. Time complexity: as above
+ \item Maximize the utilization of the EPs’ Residual Space (RS). RS is
+ roughly the same concept as EMSs in the Genetic Algorithm, namely the cubes
+ defined by projecting the EPs to the walls of a box. Pick the box with the
+ smallest RS still fitting the item. \\ Time complexity: $O(N^2 + max\{N,
+ N^2\})$ = $O(N^2)$
+
+\end{enumerate}
+
+\end{frame}
+
+
+%------------------------------------------------
+
\begin{frame}
\frametitle{Details}
-\label{appendix}
+\label{appendix1}
TODO Here come corner cases and issues. Use if questions arise
Example: why does EP heuristic mandate a decreasing order of items? Why does GA
use exactly N chromosomes and M mutations \dots
\end{frame}
-
-
%
%%------------------------------------------------
%