From 3904116595835e61dc5a051be982f73383089d57 Mon Sep 17 00:00:00 2001 From: Karine Deschinkel Date: Fri, 9 Dec 2016 15:43:30 +0100 Subject: [PATCH] =?utf8?q?version=20am=C3=A9lioree=20pour=20eviter=20copie?= MIME-Version: 1.0 Content-Type: text/plain; charset=utf8 Content-Transfer-Encoding: 8bit --- article.bib | 14 +++- article.tex | 225 +++++++++++++++++++++++++++++---------------------- biblio.bib | 12 +++ reponse2.pdf | Bin 71808 -> 71819 bytes 4 files changed, 155 insertions(+), 96 deletions(-) diff --git a/article.bib b/article.bib index b53ef03..2bb24a7 100644 --- a/article.bib +++ b/article.bib @@ -582,4 +582,16 @@ title={α-Coverage to extend network lifetime on wireless sensor networks}, publisher={Springer-Verlag}, author={Gentili, Monica and Raiconi, Andrea}, pages={157-172}, -} \ No newline at end of file +} + + +@article{idrees2015distributed, + title={Distributed lifetime coverage optimization protocol in wireless sensor networks}, + author={Idrees, Ali Kadhum and Deschinkel, Karine and Salomon, Michel and Couturier, Rapha{\"e}l}, + journal={The Journal of Supercomputing}, + volume={71}, + number={12}, + pages={4578--4593}, + year={2015}, + publisher={Springer US} +} diff --git a/article.tex b/article.tex index 629a7e4..d886626 100644 --- a/article.tex +++ b/article.tex @@ -41,7 +41,7 @@ %% for the whole article with \linenumbers. %% \usepackage{lineno} -\journal{Ad Hoc Networks} +\journal{Journal of Supercomputing} \begin{document} @@ -86,7 +86,7 @@ \author{Ali Kadhum Idrees$^{a,b}$, Karine Deschinkel$^{a}$, \\ Michel Salomon$^{a}$, and Rapha\"el Couturier $^{a}$ \\ $^{a}${\em{FEMTO-ST - Institute, UMR 6174 CNRS, \\ University Bourgogne Franche-Comt\'e, + Institute/CNRS, \\ Univ. Bourgogne Franche-Comt\'e, Belfort, France}} \\ $^{b}${\em{Department of Computer Science, University of Babylon, Babylon, Iraq}} } @@ -99,13 +99,13 @@ divided into subregions and then the MuDiLCO protocol is distributed on the sensor nodes in each subregion. The proposed MuDiLCO protocol works in periods during which sets of sensor nodes are scheduled, with one set for each round of a period, to remain active during the sensing phase and thus ensure coverage so -as to maximize the WSN lifetime. \textcolor{blue}{The decision process is +as to maximize the WSN lifetime. The decision process is carried out by a leader node, which solves an optimization problem to produce the best representative sets to be used during the rounds of the sensing phase. The optimization problem formulated as an integer program is solved to optimality through a Branch-and-Bound method for small instances. For larger instances, the best feasible solution found by the solver after a given time - limit threshold is considered.} + limit threshold is considered. Compared with some existing protocols, simulation results based on multiple criteria (energy consumption, coverage ratio, and so on) show that the proposed protocol can prolong efficiently the network lifetime and improve the coverage @@ -142,6 +142,23 @@ regions to turn-off redundant sensor nodes and thus save energy. In this paper, we concentrate on the area coverage problem, with the objective of maximizing the network lifetime by using an optimized multiround scheduling. +The MuDiLCO protocol (for Multiround Distributed Lifetime Coverage Optimization +protocol) presented in this paper is an extension of the approach introduced +in~\cite{idrees2015distributed}. +% In~\cite{idrees2015distributed}, the protocol is +%deployed over only two subregions. Simulation results have shown that it was +%more interesting to divide the area into several subregions, given the +%computation complexity. + +\textcolor{green}{ + Compared to our previous paper~\cite{idrees2015distributed}, in this one we study the +possibility of dividing the sensing phase into multiple rounds. In fact, in this paper we make a multiround optimization, while it was +a single round optimization in our previous work. The idea is + to take advantage of the pre-sensing phase to plan the sensor's activity for + several rounds instead of one, thus saving energy. In addition, when the + optimization problem becomes more complex, its resolution is stopped after a + given time threshold. In this paper we also analyse the performance of our protocol according to the number of primary points used (area coverage is replaced by the coverage of a set of particular points called primary points, see section~\ref{pp}).} + The remainder of the paper is organized as follows. The next section reviews the related works in the field. Section~\ref{pd} is devoted to the description of MuDiLCO protocol. Section~\ref{exp} introduces the experimental @@ -184,11 +201,11 @@ network. Note that centralized algorithms have the advantage of requiring very low processing power from the sensor nodes, which usually have limited processing capabilities. The main drawback of this kind of approach is its higher cost in communications, since the node that will make the decision needs -information from all the sensor nodes. \textcolor{blue}{Exact or heuristic +information from all the sensor nodes. Exact or heuristic approaches are designed to provide cover sets. Contrary to exact methods, heuristic ones can handle very large and centralized problems. They are proposed to reduce computational overhead such as energy consumption, delay, - and generally allow to increase the network lifetime.} + and generally allow to increase the network lifetime. The first algorithms proposed in the literature consider that the cover sets are disjoint: a sensor node appears in exactly one of the generated cover @@ -213,9 +230,9 @@ selection algorithm to minimize the number of active nodes so as to prolong the network lifetime. Various centralized methods based on column generation approaches have also been proposed~\cite{gentili2013,castano2013column,rossi2012exact,deschinkel2012column}. -\textcolor{blue}{In~\cite{gentili2013}, authors highlight the trade-off between +In~\cite{gentili2013}, authors highlight the trade-off between the network lifetime and the coverage percentage. They show that network - lifetime can be hugely improved by decreasing the coverage ratio.} + lifetime can be hugely improved by decreasing the coverage ratio. \subsection{Distributed approaches} @@ -267,27 +284,15 @@ synchronized and predetermined time-slot where the sensors are active or not. Indeed, each sensor maintains its own timer and its wake-up time is randomized \cite{Ye03} or regulated \cite{cardei2005maximum} over time. -The MuDiLCO protocol (for Multiround Distributed Lifetime Coverage Optimization -protocol) presented in this paper is an extension of the approach introduced -in~\cite{idrees2014coverage}. In~\cite{idrees2014coverage}, the protocol is -deployed over only two subregions. Simulation results have shown that it was -more interesting to divide the area into several subregions, given the -computation complexity. Compared to our previous paper, in this one we study the -possibility of dividing the sensing phase into multiple rounds and we also add -an improved model of energy consumption to assess the efficiency of our -approach. In fact, in this paper we make a multiround optimization, while it was -a single round optimization in our previous work. \textcolor{blue}{The idea is - to take advantage of the pre-sensing phase to plan the sensor's activity for - several rounds instead of one, thus saving energy. In addition, when the - optimization problem becomes more complex, its resolution is stopped after a - given time threshold}. - - \section{MuDiLCO protocol description} \label{pd} -\subsection{Assumptions} +\subsection{Assumptions and primary points} +\label{pp} +\textcolor{green}{Assumptions and coverage model are identical to those presented in~\cite{idrees2015distributed}.} + +\iffalse We consider a randomly and uniformly deployed network consisting of static wireless sensors. The sensors are deployed in high density to ensure initially a high coverage ratio of the interested area. We assume that all nodes are @@ -304,7 +309,11 @@ range is said to be covered by this sensor. We also assume that the communication range satisfies $R_c \geq 2R_s$. In fact, Zhang and Zhou~\cite{Zhang05} proved that if the transmission range fulfills the previous hypothesis, a complete coverage of a convex area implies connectivity among the -active nodes. +active nodes.\fi + +\textcolor{green}{We consider a scenario where sensors are deployed in high density to ensure initially +a high coverage ratio of the interested area. Each sensor has a predefined sensing range $R_s$, an initial energy supply (eventually different from each other) and is supposed to be equipped with module for locating its geographical positions. All space points within the disk centered at the sensor with the radius of the sensing +range is said to be covered by this sensor.} \indent Instead of working with the coverage area, we consider for each sensor a set of points called primary points~\cite{idrees2014coverage}. We assume that @@ -352,47 +361,30 @@ $X_{25}=( p_x + R_s * (\frac{1}{2}), p_y + R_s * (\frac{-\sqrt{3}}{2})) $. \subsection{Background idea} -\textcolor{blue}{The WSN area of interest is, at first, divided into +The WSN area of interest is, at first, divided into regular homogeneous subregions using a divide-and-conquer algorithm. Then, our protocol will be executed in a distributed way in each subregion simultaneously to schedule nodes' activities for one sensing period. Sensor nodes are assumed to be deployed almost uniformly and with high density over the region. The regular subdivision is made so that the number of hops between any pairs of sensors inside a subregion is less than or equal - to 3.} + to 3. As can be seen in Figure~\ref{fig2}, our protocol works in periods fashion, where each period is divided into 4~phases: Information~Exchange, -Leader~Election, Decision, and Sensing. Each sensing phase may be itself -divided into $T$ rounds \textcolor{blue} {of equal duration} and for each round +Leader~Election, Decision, and Sensing. \textcolor{green}{Compared to protocol DiLCO described in~\cite{idrees2015distributed}} each sensing phase is itself +divided into $T$ rounds of equal duration and for each round a set of sensors (a cover set) is responsible for the sensing task. In this way a multiround optimization process is performed during each period after Information~Exchange and Leader~Election phases, in order to produce $T$ cover -sets that will take the mission of sensing for $T$ rounds. +sets that will take the mission of sensing for $T$ rounds. \textcolor{green}{Algorithm~\ref{alg:MuDiLCO} is +executed by each sensor node~$s_j$ (with enough remaining energy) at the beginning of a period.} \begin{figure}[t!] \centering \includegraphics[width=125mm]{Modelgeneral.pdf} % 70mm \caption{The MuDiLCO protocol scheme executed on each node} \label{fig2} \end{figure} -This protocol minimizes the impact of unexpected node failure (not due to -batteries running out of energy), because it works in periods. - On the one hand, if a node failure is detected before making the decision, the - node will not participate to this phase, and, on the other hand, if the node - failure occurs after the decision, the sensing task of the network will be - temporarily affected: only during the period of sensing until a new period - starts. \textcolor{blue}{The duration of the rounds is a predefined - parameter. Round duration should be long enough to hide the system control - overhead and short enough to minimize the negative effects in case of node - failures.} - -The energy consumption and some other constraints can easily be taken into -account, since the sensors can update and then exchange their information -(including their residual energy) at the beginning of each period. However, the -pre-sensing phases (Information Exchange, Leader Election, and Decision) are -energy consuming for some nodes, even when they do not join the network to -monitor the area. - -We define two types of packets that will be used by the proposed protocol: +\textcolor{green}{As already described in~\cite{idrees2015distributed}}, two types of packets are used by the proposed protocol: \begin{enumerate}[(a)] \item INFO packet: such a packet will be sent by each sensor node to all the nodes inside a subregion for information exchange. @@ -411,43 +403,68 @@ There are five status for each sensor node in the network: \item COMMUNICATION: sensor node is transmitting or receiving packet. \end{enumerate} -Below, we describe each phase in more details. -\subsection{Information Exchange Phase} -Each sensor node $j$ sends its position, remaining energy $RE_j$, and the number + +This protocol minimizes the impact of unexpected node failure (not due to +batteries running out of energy), because it works in periods. + On the one hand, if a node failure is detected before making the decision, the + node will not participate to this phase, and, on the other hand, if the node + failure occurs after the decision, the sensing task of the network will be + temporarily affected: only during the period of sensing until a new period + starts. The duration of the rounds is a predefined + parameter. Round duration should be long enough to hide the system control + overhead and short enough to minimize the negative effects in case of node + failures. + +The energy consumption and some other constraints can easily be taken into +account, since the sensors can update and then exchange their information +(including their residual energy) at the beginning of each period. However, the +pre-sensing phases (Information Exchange, Leader Election, and Decision) are +energy consuming for some nodes, even when they do not join the network to +monitor the area. + + + + +At the beginning of each period, each sensor wich has enough remaining energy ($RE_j$) to be alive during at least one round ( $E_{R}$ is the amount of energy +required to be alive during one round) sends (line 3 of algorithm~\ref{alg:MuDiLCO}) its position, remaining energy $RE_j$, and the number of neighbors $NBR_j$ to all wireless sensor nodes in its subregion by using an INFO packet (containing information on position coordinates, current remaining energy, sensor node ID, number of its one-hop live neighbors) and then waits for -packets sent by other nodes. After that, each node will have information about -all the sensor nodes in the subregion. In our model, the remaining energy -corresponds to the time that a sensor can live in the active mode. - -\subsection{Leader Election phase} +packets sent by other nodes (line 4). -This step consists in choosing the Wireless Sensor Node Leader (WSNL), which -will be responsible for executing the coverage algorithm. Each subregion in the -area of interest will select its own WSNL independently for each period. All -the sensor nodes cooperate to elect a WSNL. The nodes in the same subregion -will select the leader based on the received information from all other nodes in +After that, each node will have information about +all the sensor nodes in the subregion. + The nodes in the same subregion +will select (line 5) a Wireless Sensor Node Leader (WSNL) based on the received information from all other nodes in the same subregion. The selection criteria are, in order of importance: larger number of neighbors, larger remaining energy, and then in case of equality, larger index. Observations on previous simulations suggest to use the number of one-hop neighbors as the primary criterion to reduce energy consumption due to -the communications. +the communications.\\ + + -\subsection{Decision phase} -\label{decision} -Each WSNL will \textcolor{blue}{solve an integer program to select which cover - sets will be activated in the following sensing phase to cover the subregion - to which it belongs. $T$ cover sets will be produced, one for each round. The - WSNL will send an Active-Sleep packet to each sensor in the subregion based on - the algorithm's results, indicating if the sensor should be active or not in - each round of the sensing phase.} +%Each WSNL will solve an integer program to select which cover +% sets will be activated in the following sensing phase to cover the subregion +% to which it belongs. $T$ cover sets will be produced, one for each round. The +% WSNL will send an Active-Sleep packet to each sensor in the subregion based on +% the algorithm's results, indicating if the sensor should be active or not in +% each round of the sensing phase. +\subsection{Multiround Optimization model} +\label{mom} +As shown in Algorithm~\ref{alg:MuDiLCO} at line 8, the leader (WNSL) will execute an optimization +algorithm based on an integer program. to select which cover +sets will be activated in the following sensing phase to cover the subregion +to which it belongs. $T$ cover sets will be produced, one for each round. The +WSNL will send an Active-Sleep packet to each sensor in the subregion based on +the algorithm's results (line 10), indicating if the sensor should be active or not in +each round of the sensing phase. -As shown in Algorithm~\ref{alg:MuDiLCO}, the leader will execute an optimization -algorithm based on an integer program. The integer program is based on the model + +The integer program is based on the model proposed by \cite{pedraza2006} with some modifications, where the objective is to find a maximum number of disjoint cover sets. To fulfill this goal, the authors proposed an integer program which forces undercoverage and overcoverage @@ -457,7 +474,8 @@ consider binary variables $X_{t,j}$ to determine the possibility of activating sensor $j$ during round $t$ of a given sensing phase. We also consider primary points as targets. The set of primary points is denoted by $P$ and the set of sensors by $J$. Only sensors able to be alive during at least one round are -involved in the integer program. +involved in the integer program. \textcolor{green}{Note that the proposed integer program is an extension of that formulated in~\cite{idrees2015distributed}, +variables are now indexed in addition with the number of round $t$.} For a primary point $p$, let $\alpha_{j,p}$ denote the indicator function of whether the point $p$ is covered, that is: @@ -556,14 +574,15 @@ to guarantee that the maximum number of points are covered during each round. In our simulations, priority is given to the coverage by choosing $W_{U}$ very large compared to $W_{\theta}$. -\textcolor{blue}{The size of the problem depends on the number of variables and +The size of the problem depends on the number of variables and constraints. The number of variables is linked to the number of alive sensors $A \subseteq J$, the number of rounds $T$, and the number of primary points $P$. Thus the integer program contains $A*T$ variables of type $X_{t,j}$, $P*T$ overcoverage variables and $P*T$ undercoverage variables. The number of constraints is equal to $P*T$ (for constraints (\ref{eq16})) $+$ $A$ (for - constraints (\ref{eq144})).} + constraints (\ref{eq144})). +\iffalse \subsection{Sensing phase} The sensing phase consists of $T$ rounds. Each sensor node in the subregion will @@ -571,6 +590,7 @@ receive an Active-Sleep packet from WSNL, informing it to stay awake or to go to sleep for each round of the sensing phase. Algorithm~\ref{alg:MuDiLCO}, which will be executed by each sensor node~$s_j$ at the beginning of a period, explains how the Active-Sleep packet is obtained. +\fi \begin{algorithm}[h!] \BlankLine @@ -636,7 +656,7 @@ $W_{U}$ & $|P|^2$ \\ \label{table3} \end{table} -\textcolor{blue}{Our protocol is declined into four versions: MuDiLCO-1, +Our protocol is declined into four versions: MuDiLCO-1, MuDiLCO-3, MuDiLCO-5, and MuDiLCO-7, corresponding respectively to $T=1,3,5,7$ ($T$ the number of rounds in one sensing period). Since the time resolution may be prohibitive when the size of the problem increases, a time limit @@ -653,7 +673,7 @@ $W_{U}$ & $|P|^2$ \\ size. After that, this threshold value is increased if necessary so that the solver is able to deliver a feasible solution within the time limit. In fact, selecting the optimal values for the time limits will be investigated in the - future.} + future. In the following, we will make comparisons with two other methods. The first method, called DESK and proposed by \cite{ChinhVu}, is a full distributed @@ -669,8 +689,13 @@ lifetime. Moreover, it makes the MuDiLCO protocol more robust against random network disconnection due to node failures. However, too many subdivisions reduce the advantage of the optimization. In fact, there is a balance between the benefit from the optimization and the execution time needed to solve it. In -the following we have set the number of subregions to 16. +the following we have set the number of subregions to 16 \textcolor{green}{as recommended in~\cite{idrees2015distributed}}. +\subsection{Energy model} +\textcolor{green}{The energy consumption model is detailed in~\cite{}. It is based on the model proposed by~\cite{ChinhVu}. We refer to the sensor node Medusa~II which +uses an Atmels AVR ATmega103L microcontroller~\cite{raghunathan2002energy} to use numerical values.} +\textcolor{red}{Est-ce qu'il faut en ecrire plus et redonner le tableau de valeurs?} +\iffalse \subsection{Energy model} We use an energy consumption model proposed by~\cite{ChinhVu} and based on @@ -727,10 +752,12 @@ stay alive during one round. This value has been computed by multiplying the energy consumed in active state (9.72 mW) by the time in second for one round (3600 seconds). According to the interval of initial energy, a sensor may be alive during at most 20 rounds. +\fi \subsection{Metrics} -To evaluate our approach we consider the following performance metrics: +\textcolor{green} {To evaluate our approach we consider the performance metrics detailed in~\cite{idrees2015distributed} which are Coverage Ratio, Network Lifetime and Energy Consumption. +Compared to the previous definitions, formulations of Coverage Ratio and Energy Consumption are enriched with the index of round $t$.} \begin{enumerate}[i] @@ -793,7 +820,11 @@ indicate the energy consumed by the whole network in round $t$. %\item {Network Lifetime:} we have defined the network lifetime as the time until all %nodes have been drained of their energy or each sensor network monitoring an area has become disconnected. +\end{enumerate} +\iffalse +\begin{enumerate} + \setcounter{5} \item {{\bf Execution Time}:} a sensor node has limited energy resources and computing power, therefore it is important that the proposed algorithm has the shortest possible execution time. The energy of a sensor node must be mainly @@ -805,6 +836,7 @@ indicate the energy consumed by the whole network in round $t$. to network disconnections and for which round it occurs. \end{enumerate} +\fi \section{Experimental results and analysis} \label{analysis} @@ -812,12 +844,12 @@ indicate the energy consumed by the whole network in round $t$. \subsection{Performance analysis for different number of primary points} \label{ch4:sec:04:06} -In this section, we study the performance of MuDiLCO-1 approach for different +In this section, we study the performance of MuDiLCO-1 approach (with only one round as in~\cite{idrees2015distributed}) for different numbers of primary points. The objective of this comparison is to select the suitable number of primary points to be used by a MuDiLCO protocol. In this comparison, MuDiLCO-1 protocol is used with five primary point models, each model corresponding to a number of primary points, which are called Model-5 (it -uses 5 primary points), Model-9, Model-13, Model-17, and Model-21. +uses 5 primary points), Model-9, Model-13, Model-17, and Model-21. \textcolor{green}{Note that results presented in~\cite{idrees2015distributed} corresponds to Model-13 (13 primary points)}. \subsubsection{Coverage ratio} @@ -876,9 +908,9 @@ greater than 50\% for far more rounds. Overall, the proposed sensor activity scheduling based on optimization in MuDiLCO maintains higher coverage ratios of the area of interest for a larger number of rounds. It also means that MuDiLCO saves more energy, with less dead nodes, at most for several rounds, and thus -should extend the network lifetime. \textcolor{blue}{MuDiLCO-7 seems to have +should extend the network lifetime. MuDiLCO-7 seems to have most of the time the best coverage ratio up to round~80, after that MuDiLCO-5 is - slightly better.} + slightly better. \begin{figure}[ht!] \centering @@ -907,7 +939,10 @@ efficient manner. \end{figure} \subsection{Stopped simulation runs} - +A simulation ends when the sensor network + becomes disconnected (some nodes are dead and are not able to send information + to the base station). We report the number of simulations that are stopped due + to network disconnections and for which round it occurs. Figure~\ref{fig6} reports the cumulative percentage of stopped simulations runs per round for 150 deployed nodes. This figure gives the breakpoint for each method. DESK stops first, after approximately 45~rounds, because it consumes @@ -948,12 +983,12 @@ consumption point of view. The other approaches have a high energy consumption due to activating a larger number of redundant nodes as well as the energy consumed during the different status of the sensor node. -\textcolor{blue}{Energy consumption increases with the size of the networks and +Energy consumption increases with the size of the networks and the number of rounds. The curve Unlimited-MuDiLCO-7 shows that energy consumption due to the time spent to optimally solve the integer program increases drastically with the size of the network. When the resolution time is limited for large network sizes, the energy consumption remains of the same - order whatever the MuDiLCO version. As can be seen with MuDiLCO-7.} + order whatever the MuDiLCO version. As can be seen with MuDiLCO-7. \subsection{Execution time} \label{et} @@ -980,9 +1015,9 @@ for different network sizes. \end{figure} As expected, the execution time increases with the number of rounds $T$ taken -into account to schedule the sensing phase. \textcolor{blue}{Obviously, the +into account to schedule the sensing phase. Obviously, the number of variables and constraints of the integer program increases with $T$, - as explained in section~\ref{decision}, the times obtained for $T=1,3$ or + as explained in section~\ref{mom}, the times obtained for $T=1,3$ or $5$ seem bearable. But for $T=7$, without any limitation of the time, they become quickly unsuitable for a sensor node, especially when the sensor network size increases as demonstrated by Unlimited-MuDiLCO-7. Notice that @@ -992,7 +1027,7 @@ into account to schedule the sensing phase. \textcolor{blue}{Obviously, the pre-sensing phases, on the other hand a leader node may waste a considerable amount of energy to solve the optimization problem. Thus, limiting the time resolution for large instances allows to reduce the energy consumption without - any impact on the coverage quality.} + any impact on the coverage quality. \subsection{Network lifetime} @@ -1011,14 +1046,14 @@ lifetime for a coverage over 95\%, and a network of 250~nodes, is greater than %This point was already noticed in subsection \ref{subsec:EC} devoted to the %energy consumption, since network lifetime and energy consumption are directly %linked. -\textcolor{blue}{Overall, it clearly appears that computing a scheduling for +Overall, it clearly appears that computing a scheduling for several rounds is possible and relevant, providing that the execution time to solve the optimization problem for large instances is limited. Notice that rather than limiting the execution time, similar results might be obtained by replacing the computation of the exact solution with the finding of a suboptimal one using a heuristic approach. For our simulation setup and considering the different metrics, MuDiLCO-5 seems to be the best suited - method compared to MuDiLCO-7.} + method compared to MuDiLCO-7. \begin{figure}[t!] \centering @@ -1053,11 +1088,11 @@ lifetime, coverage ratio, active sensors ratio, energy consumption, execution time. Indeed, when dealing with large wireless sensor networks, a distributed approach, like the one we propose, allows to reduce the difficulty of a single global optimization problem by partitioning it in many smaller problems, one per -subregion, that can be solved more easily. \textcolor{blue}{ Furthermore, +subregion, that can be solved more easily. Furthermore, results also show that to plan the activity of sensors for large network sizes, an approach to obtain a near optimal solution is needed. Indeed, an exact resolution of the resulting optimization problem leads to prohibitive - computation times and thus to an excessive energy consumption.} + computation times and thus to an excessive energy consumption. %In future work, we plan to study and propose adjustable sensing range coverage optimization protocol, which computes all active sensor schedules in one time, by using %optimization methods. This protocol can prolong the network lifetime by minimizing the number of the active sensor nodes near the borders by optimizing the sensing range of sensor nodes. diff --git a/biblio.bib b/biblio.bib index bed9b2f..d5643ba 100755 --- a/biblio.bib +++ b/biblio.bib @@ -469,6 +469,18 @@ ISSN={1536-1276}, year={2014} } +@article{idrees2015distributed, + title={Distributed lifetime coverage optimization protocol in wireless sensor networks}, + author={Idrees, Ali Kadhum and Deschinkel, Karine and Salomon, Michel and Couturier, Rapha{\"e}l}, + journal={The Journal of Supercomputing}, + volume={71}, + number={12}, + pages={4578--4593}, + year={2015}, + publisher={Springer US} +} + + @inproceedings{xu2001geography, title={Geography-informed energy conservation for ad hoc routing}, author={Xu, Ya and Heidemann, John and Estrin, Deborah}, diff --git a/reponse2.pdf b/reponse2.pdf index 1adc53501057add1ee91f285aec212c67814404e..e87ad0aba56fe7f95efe2fa0dfdee22656d32fd8 100644 GIT binary patch delta 17990 zcmV($K;yrFu>^~;1dtX1Adwer1VSJPl97iw0k4t47JuaP?frTg7T)9Pc}?dML zjO6uS@1Kas@-QLl=HcntstgvZF!alm_y5|cWt4hQ6U9Suuas2^6L8@4-{DR-&ue3fI5Um`2X7Wq^jA?!(5Bl|%#lb2YiC^l_6n z?~3OMQh1R#_KQ5GeIFZIXbxEQoZTaW*G$@yfrg_RXb_ak@enaya&Ig*nczku>}eBuP}Hb+%nGmG(Ow$w@w?q z)io_qeDrV21U@=h%f2Vhvut{Ip9g7PmSN!O?#~`h%+FGkeh6A;+Bb}=kF_G=7f~IC zBjdN~r0lQUM5i3&KFNN2P|csXhpbuPUI7aEn9ritN8Rv@aF%)!?ZKWUDk48C=IBX- zcYn|$oYosXn4wd$Bf+T1{5(&9$o-;Z;ZIVq5pFAGz#w3!D|QRN1Ye!u`{T|?6pvF+ z;E;J+(aGIV6R(q^Qq7)b^p2RX=z^B2MQ>ozdVX{)a zXj`osf(}U8H-Q;MCA;S2(>~m;zu|t9I|CVi!zu7xQRE=dW@HuQgkEpp0q}G< zZTCjPp1LQX;CI{Ut-UoaQOtant&ztKqx|f6kb)p{#6d|;g@oQ10iJ2U_@2n#>f=#1 zre~oYk4^nb)Z@b~tpH47rcriyL3hAudm0$O3~m8zoq`dPBJ=K?+J4yYWk1HRVD4R@ z42=$dcVfUGu;Hc0 zCacN97s0xb-I(iB|83l0u{+chgkEd68)`9nd&n1o$vG;J(;MBk_B4UJ0}!3;bv-a- z;oJK1RN&fX_!U@cb600kJJkL7{cy*d{#K^q}jCCgibI9k@+x z-`eKM??Dg|s9>J3meSR5{tWObUSdR{OQ^`+NDPrk?WE|{re+1>jW4{xj_(*J913IS zy$P;kEREqc_e4*V%8;}T7>>yhl{1ZrXO@UVqThHxntDD0JcZx{;Z2g-qo{ZRXMlu% zujlk>A~}VCEiYV7RFG6{pbK5y0R#kvK{B%wjAo$iX{Umq8Jvng$aO4t>II=o13l6(acSupjW!CrZ4Gya>R|Uu-@QJv zghet+9++YCTu)--J%MlUITd{KB=p*UaR@;a0t8E)cA^krcb7I;@cFEAdFdfkhbf1O zZ1@ok7oDz#8k(JjkzzQ23TPa{N#k&)Fmy80<{}LIvM4~*imdJRvcwYu_-=;}4li*Q zCt%u!74k;>G=j|D&?lHs(@K2O!5$o?obprT$1nQ^AbVtGtuw72SYmE#!Ulx`5;~L7 z1tn5Z#_H@{5zc&k-cth>8B58&jay0#C|4<{W9dkkFS{f zWgdhmmirk@^8-CBR~ckr_zl7JS6S@cTF%pj=a0kbN|Q(iBn^(21C-Qr6R}u-W)j?! zc?KST(B>eFqwI8kf-rK=@&Rz5w4uvaHYeapO(Oyk$4P+#UyzN4y1J|_XFTDYYE}$R zI`s^kz3jMd4V{-1E{_%=2Sd&q1qKakNEGIgUIhq%w+n{iwLjCd6+KhmgEUaS^jV?x zF%+Pyv8s(8%$bI2=)`;J&pPb1DYX@P4_k$Q|G&tOU{OZ#D@i7pPLdJR_%2CGMHsN3 z$Myw2Vm~5+fJPK6vz}y4KneuAmYuZ5f>Hi8*oXvVH84V)157hm=URij;Pjbudj_NY zr>(h&iSpc?B`+LMfP%;jy%aT#@aa&C^oF(cl2?rhm(^*fj0MeW3(?>`1i%YGUpQ%h zW*|}{+w4_J=kT}H+2G>CD!!0j$p z&Ul6?MmY=xGZ@D2(V2hdxkI2P5;l?Y#W`zAU1ty zYOE^(#NQ6f)Uz@id@}3~ZM4s@C!!fE)M`wXPNM^dyVXo9f4eCxmxjcQ%d zR`YgXmy3-L7E5plfLo8nKb2+?cZ^8A9BfiA9lC(^?fHQSyBUm?Y-^3%y}I>()S8!( zTNHb!y|Z;Qav{HCVHbdYWez0KjONA@*?$9iY(3-5yB^HNvfBNTlX#R6sVH|9eslU4 zff)RU!9?VADrhso6p8c?>uGKpRTMkbiWZ{7&|5FRO;05N+4RR8`HVA(T!0!;pU z*aHI7nQMhv>BkuaxHR*FkO3GA&SY-aAN~hDd>BxZ9tj!&AhS6M>`}zORdEc{hw%^s)=X-sw@A-|B-N0Bt z(H?0BQ$r%K0$@Rq6hK*5R}%~ZfIz~6AdnClC#MM<>k9i@PsV8mLu24bgw%g{D1W12 z5G+oo0>R=Ob&&{w)&o}nSQr2nlLCuNfj|Hu5J>WGMa`@am65U@qZ8x2;3E7 zhjaMVIs~AmXb6Dd5dOi3fui9kEJhFmcm0VG_%jXem{k$>%1Adi7y^qS``Mog91Vlw z&fOdMXSgm1q$k4nZ^;3Uuy^;p+Ww_g^Ch z>KU7wDVy{Cf&4F{k`mGj;42_31P~As1p&Yy5ea}e?!o_`dH!_*e}AL@r8R)S|8(~s z88s0ONWjk$KT+V0?{6d?f2hIzXK#1_|16=0#1R?>;QlxN?}J1^P}~>z|GWjiLjHf` z|C!}~<^O+or1rqo^_QLdx5NKqhq%FAz5jUN2>k$yTMk_$ZbcCP?P?DDz0tZbd-wyl z|F&vkA-E+`L^$GzE`I0UQf;`c0<4r-7^C3P->UkQn&Siw744 z0{xc_w|h_*+~tD7&Fr@cg262!_SblS$}rrj{j(2M1Qcoib0>vF#Q+dA8sbfcqXbS8 z1^9w-%WDtw`o(zwP!NH{;z9s8%>DoeB%18!pd~;6putboFMkOPmIMGje~JI70)2o+ z%fGGx7|aU>C7YT+LZw5Tt3z6jDixVL1y+Y-h_=&?EQADx3av@7-Krxt zO9{r=S<~9j)uy_@YSzwedP?p`IrEddzAH6swca8J0&HsQj%;>u6PE{9#u9xGnQu=} zHsU=sPf=o(Dt|$D;~B6-Q?--I+dFm}b5i>KRo(e9(UXg|R83qfv_-GAy|Oe*hmKj& z?8_%3Y?ED|$*Slc{a?AC+@1t4$J~!h)yuf8Ez^2Tmf%@IwbliJ0d zY?B;fz+ZfhS*TI7v`T5{yn?kT`n6xvl&yU6uj70j>){ZF-~3l)3iXOAA6*vDroTgQeh-Ih_Ur+7 z=$N^XrGG@4Q;UgQ@vLM>bD#yaurP%OJgwaHIV}P*KB+@8g#epKJ~nY>*t} zdEA?xsT5}XcmX5){>c-UJbvpKyM&DIJ_Ai85b}Xg5gR4q<;@#JG_$8YLnq*EzJK9q z>d`BdU(D;Oh5Av*ES1x&7m)<5UUK~olNvQAmTXogss|m;8Ag3k0YJV}@|6|3O+ZE( z)tsxuijT!?0;+m1H#1v<-Ry8S$iAMvK2X|TB)`CneWcp zXOBlYTpC>u<~eQl+&HYL(j}J#x`OjoZ@EBp1|Blb9;cbSHvs#72)e;Yi^z+7Jf5RA z?Q3W9da%WTIPBBcyqCU_IwM_dJt z^2``rk|GOd&4e{=lL2=~Bb(P*m^heS|oelqLInM)CxH#>sz-h1vgW1wqJW83->kmA7Vz-&=Np?kto^w|weeKL2|hke0EmryfJhDri4ZhsrJ?(L&UY^Z%; zVDQ{?y{#}ID<~YYbRe>ODZ9<`4xRqW!+WN}$(pxALwl;q@BZG$yOEwZ>KJvm3WWK2 z>$;*mMEwPy7fBZ1s?5DD25G6Kz7D&)k>MXlAofhiUppA|&Bm;fZs=9a)tNzHiCBBU zKqP3#MJXYo=fd@yoPTG+yv3%>Cyl$Yl|6)B9yG6S_y?}232kmNgQ^4;bncqndv5nd#3ILX5cFWs znNCi1%%wW8n%$ziT_>)!{^)6MylHCXLFO_*H{A8aM1#2xF>>LZ*#$H0>A5_^genrw zWf@$hg}N|7pd`L@@37{T&G%|PNgTzk zfuYdBKrFs=q z8^2B?xbI!ZgSaK%oMX2BNUNTK^_ZBCt&hnC5}Id8=YKOh=#W+wNRvA0pvgOD>A)b1 zuS-KGJj%n_^5a=KRigXz@_K!4qPF`2_t;|?SaY_tI>|J6n9+s@adj+)q6Y;JP3Y}( z{DE>qqAoOF=nrKLJ@fT1#OXxOBJMCrB~S9_o7&Ln7|6%h_4eP=Q!ZaK);$=9Z;w+< z%Ozi@?tgiI7_oY7^@lL;r1Kr~&TN%UiS=lXol(D?9QFpasD=K)sKyJ%jcJ`kX`6>+ z_J%21njH1bDV=xwxIfalj$ z8M84zkv#Z`d@u?5?u6o4_ky=8_oHQ2~W4Ha-?lr==v*^J0H8yvX{ z9t<|AN;cCu9})F7*Y@)X9NAq?{q8qikG)#xtYp6tY9?FmGbKFTZZe@2(6VM(-CB1s zE`Mx$?vZzaLZE|9uj!hZKhHc6{CwlC0Cqh+ZKLvKwUAHt_>cIIBD`=i8DlXlxX!g} z&i0jE##$6!ZZn?d7Tqw54NvyH^xZKV7?tXEV=P6HDOB{!JLM&X=tszIQL0`q9hG-& zifK{x56`>xidjAwkpOg}zh*h~c&!PkaDS&f8bst(Pfujcz-MEhzaM@UgQ4i0YKaa0 zbRvkJC#fW0-Bq=GCO}Yj-tcAr`g(9_kk9Ls)hFn3BFfFh5so6I=j1y0z^J&-QjJP< z`+$l0Nv=S3rks%>@B=ZB6~;D*EMYI%FQYk;yNXUEccFxPr zeBa(NH!@G%YrH^l0JwfoK}slVIaHq~sA>1IY69VF6`&Z z=~97`8=q~41YY^O1n3MaeEwSYhGIZnw#)j3)5lF7mx|0LIH`o)X#9}^#`@Ngm-?k( zd>9Gk7QOyqSjN5RZ_X9VYJapQvn?Me_&Ffk!Hh;Tl_wa7q}qbQkN4dIp)O!#72n!J zlJLQ`1;W4%)y;6b?Pz}Ga`)X%uCQ~98(W$MnLT4^gy`4zXX@WPdVBfl)Ch*O^+;H1 zM%&*)$$xo2R$h_iFs?jwlep^HM3Xtu0V-^4*M&Ggc&JY?kBy3Z%703Cg<+DKl9Y<} zU^dP==;iRZesY!T-1`sSvdnv>);^-DF+mJI92buUpkL=f11`o)WXR=MsURIEdW=})jSI|j>n50@6VadNx z`m|O2#mL)nXeu2hpnp2dxEh*>d}U(yX|7FlrJ|wj?6pu*tK5q(+FtI$okfdpZLlRQ zFEcmdKKKorPjW_p3I=7@RO1yYnv(>#g{&9RHAPan+$#jJ$JqU4KmZnYg;+hU}e(l1J?sb%)13k?E?kkOGBi1h4TvF zmHMJsFL-5?*>m@6@q>*OFOptTO`twI7V2H`^tM`)xi(Va?PO7rK!6gwzRtZ*UfJs8TKSlF5 z%}1wSGlk|D%I{DVg%hSiQ#@Uc3ux3k&gU zu`n3RGke|5z2P%Iyy@oa@%nM69+o$E@J9xe3V%-}y=ss#f!szK>oPVkkeR(H!opcn zQct8-dc2*^aP4UBYc=B)0!95??(#6qTZbx6?%7k+kjjk4Z6PkAGyzl;x%PHIWu)Em zo-n)Y1Jzx_jigi(?kyD==^E$CuKtFTg(=O|V<8MJOMd;rsbZ~Tn0n$nJ>%XeXphj? zS%32+nno~eZdrU>`LVUKQr=SZdL^yw96D~;Hf4u4%hmlQw~*CbTe=>MDO-Cz2~@-K zMZRPN1x`8qL2ezJwsW2KCbg|Wfbk`ptoGccyz23qHdAIMR}J58WC8j%VCY%K}AyTgmgW9PTtycwHIpAGIuX?KZKqIbo%YQXjmW}qrS0c0z|i+ z7qeAtD|Xp!w)7U%p02o-ak~2KfifTgn9qqS2FiLO+V|_{cGl!o@BA=Aj2tH4-du9I zGnc!UpkSPL;-^Ku65%yum~gCX>K-h4N8@Ek(1`$bS^O#++V14a+KMsu5aAQWV<~P7Nx1a0!t>XO zW$ccVvvQlRkS*uZNAR&q%Ti<#s?R)p=p#WTZj458g;jGzy6msKQ&=XcD}PRlhUf-H zyh*u5d#Aq|=69O5XIJrwHJ$t#RwrCyxC)foAK~|YbhA~Tm1n{I!zz_pGL|$&KNr71 z)~mXxxtl0@n&M-z&04qYWg^u%UuwmOzT|3{qW>WM)BEI0-qw`-vR*bZ9WP9YUOj1F zcq*n(67H-ty(}RgDj)KBRey!(RV-dSv*qJGZT722Sz`zDcIyn&5X;qXsW#h<|IYqaaFLri+?#veAV6CXhf!$ zb;*ldc26M7BZBZu{XMOJ<=v)lKG5{By^8k&JaGIUZe|k;++{i$Da+=L(a)}2poe6E zRg`U9U^NqeRjVH@!cP}Uj5(RPbQ`^63+^fxCXw15FEqaN7(}v=XJm!3E=h#cTRexSOSwCzH6F+n$5k%7s za&6eItbK=(F>Ch3E9(W1Bpp6w<7IoI#t^)dDp1-vQ*%uH&<80R+*dUFqQ<_AfJ*gY z;%PruC+(w;#62a2M-O$H1m!`sa{}Yb)5f;S*YF;RyTeQ}3h7E@?25=yC~S4Z{75C% z#9ZGtjCtg{h<~8I7l&7Z*JtVGoNYS@D&kb6GG=E2_w2jEh@ z#O5u0JA|p)H7#vt6DEY7pPyme!%fOsogRsyr#E+W61k+z??!dCq{jo69iJE0>DbG~ zM26#QU(KXYI29hh|IJuWO5$k&-CvG`8^sGGrtE!nfb{cPFIv^f8y6EVKg~q9}cGjde-ze8?kO-GvHd1Ph6%ID!g$K}VHsCI|x z1$+d-1zxLY517)vtgtEmt9I8GjinBmccVLLtx|hEN2kJH3=#ce8|>y=i)2G_C#pB= zdo-K2#g6z-HBCA=!ecLnyfdf5za?JLNI}7y)_*R|piVhf^sZbp3s2QL=cJn6Q0W(L zaNOyOB4JmND#eSY#tY@-!M{oip%DEGlX@c0{KlS!_e3gu0<-a3RW;@W2wJp_84 zcYE-l+$XQx=_q*W$TEnlxSK1y{c>&W%w4cNy8no!ZSF%+_^qy?qOD%nBr`AJeUt3x z*A6&!?0guB?Tmd!Yh{p9apT6X5dk5$;D4Te27A{CftN}ZHps?i_00&h~RY-UNg^*A6P^14gW?^P9stR19~ zvg?Q^#~4Wxs|jEy84<)}BpVE(x9e(iHX|oRMw>UmaB(q8%s9#r}X>Oz_SE`#1IiBnykZu__ z=y{;O`k9VEiQ4vEw25WibcA_-Z-K3(_|ww)Gn=(~hS;~ox>wqPn^sRbA}gjU6FR*c z6td?C-w0=MadnIyz7g`0;@5;O9U`?imMYXf6VK$mB0(2?HNTN)Z?bu#XxdM^kzJUbWiIl+} z`8w_TgZ;N=k(RW|)3rKoBfMHA4>Y5$+Sw|YM7%Cy#PK?BpHRW( zx48$0sFVva1AU3gcV-2T&-&Xx8WFoYG6o6lSE`QZ35MHSoDy6Uv2B6eF1nRk*W!pi z-tF!mOek+jx`H3YR;Yb1wtp)s-+K$}arHx3X;cEsgt5;yU9)l}xSyCBcxoz8v!+Kf zpp1E1EFpZW;PF;phuEpsw?h{l#y4%T91NL6=#$LbUQBr#e|=nxdmNq4KHHN)d8;P> zAP#bO{~7*a^cQ>DD(-_RJlL7E_QXSGhDi2{#c_5fN=fHuE2RGepaZoSfs@fX8WA8o zJPI#NWo~D5XfYr$G%_@k{5mB8HO7ID7ZnvNR&Z_|4Q6MLM3wpb2w;bB06=15BHX{j0dmezHy8vA2WWzk_E2Y3 zMF`jlppSsSph%y8rC^t|MZ=pSUfcq{+|dzkwly*|Pg=>>L!0#FJk7z7G;M}I|lz-^#z090{+zN!X5+XV{$ z~!yMW<7Ft{DS7Ul#6Xe(>*BE6B^05II< zmm}E89f67mdxBw3U~81a@6N#hWjS2{7zOYzciLx z{)7K-?7y`-VAx;10sl#;3b#c7K)^p3P?z^N4$r@6!T#6EZ~*>YLJNW7G8Dl65B1G~ zLVrLA>I3@!UVz^r|2N+M%JToB{(m>5?BV3}+syvk{(qRk&M+sRzZ_6h_CTW6LKA^n z4)}k&8bklAvL@69=HdL`R#hYzwH|VCI~2uvK!UtL!9RMKyE4ohYNG=~LhSz#>F-&f zYB<5*P#uIj?AMEfiUI=v%ZA!Ih$HG%aeqfm?2ifTj#@$F@8SNEp{OPMcOOb{2*T#q zHu4Jz1Hf)>U>`gbAyASKzz>94TN|kNZ@vTgc;N^nDg=PS900IIxZ(X8v#1b&&+wP( zw*(Re@S)~|1pn6js}RJ)%?;JbZ+@V<_*?${20)?SPzc`a6apd<>QEcnc2+G%($C4BQ!i7V?FA`YICN^syPm!LI(_7~T1Q{+BY46? zul(JP-a2vW_5^)0+3%E=af+x3?X_{L{6q0#w>~r-x|sX#?#kHM_35(^`Losb%p`lU zl9niosD-s@b-b%C$<`pJ6*Jmd(SJiWTMrnl2^~|i<*qm>kVklxTkN@=`M}Rwg+%@lhuH*2 za!KM^R-f@UR$xTcp!0g%x<|f5caXz&*;TgOm(Y4TVIr4oE8XGuYd`90j(>s|)Q;(I z2PrXw@3bXU3hc|ie7I9^_^1s?TjkVSP6#kup_S6VfP=)3Gw)8*`x^4cGrXA|tioFx zZl>!;FWey(Uf5&ZHDR!MVOI7urpGS%OZCKgVRV|^_Tw7^8w#2F08Yv%`76gyB?Bzd zLE{QTkg#E%NVz;s&6$=0(SJLZaiq-hFeuSfOOt zM|0h4IZpA^_6vQ^Hs^)IDQU&|bAneZr(Y;u~j@+!*qRVKz4UhaI+^Ory|ScHFH>Gv)Z5qcfH0&%uay(Bwil8fC@Nm8ku zgv!W+4%^rZLh^^iaMv-fIIN9*8QrfQ zABj~Bh$OO}>f*d5qh`gpSs(MHJ-gzwOy@E(%8MZnST>eoXetAQf^19?f=7< z-|Y*X?$qih`5wrHUHXL7%cNT>*>C0#;_WN|x1+R5f=6g7VwCIrnRDoi40505PU@km zsc0gG=w3qhv_o8mR$7bHm!$`Zk`G<|W>(*y2OSp22k-KwM$$eY!%7S#O;NwR-WKik z8Qhv-oO^Hh`hVswS5|+@oOnfhXkeZhsSMf`akl>lX9>wgO}@D)JMi3@4(YEablW!Lu1Y7p8ZuuqK%1jtBjyugi6a$%RGgb18$HbRmM3cAs44`X2Q1Bt_UXMs~+-* z?d#*|zREl!P7>;D3!cZ1?4Kl_Uq@>8*o?7Od)XqMtu{Qav-Hk*tDV>I=}ZM-l)TIW zn-=x8dVfhO*&p5(PNKg=(ZBD<3PCo8ni*ndX#-j+NIAy|uPF6Mo6g|~6>7hxaw zZG{IuJ2qhveNw84GhUwC#wcz9je74K!_VOv2bd=P`3_OUGSv|I`8Uh2(toz>VN&i3 zYI(&|wk?Yw1$n$*Xsi44Qg^F8N@E$x5ax9_4SNK`D!bl)E$w9YgY{uViBQ$n_1&!P9S&6&2nif>wia zh@R2cPF0;LdNW#grgJE$zBlwdy5=J9V&N%D4o_=6G0K^6BaA71ZOmzd1usQrf38?w zr>0bN2m>wAVP>^&yWLREmq@`0YAQ0|AAfm?OEp{C&~a0$nB&#-GfES+?DLkdstDfO z%=Twuq@4_yC~!ydF_R?F)9*V{w&Rb9DGQDEve3w_y;}X&w z^>TRx^7bCrWPEXaRQc1+Bf@VMf&{JH6zQXkp-ky;hKBZ^g+m6esvQc7q6-hV|Q~D0v1OO|qVLb@ge?*9`S~Q|IP+_jyjlw#l<3 z=)y|T`*rb%yqZHIHCP;~B$sV5Cx4ug)ZFRQ614mBAv#inMJmISf^xvG#8CpOk24HO zFh8Hk?5Hf*PJ8Oc5Xzp-CGTkgCR6OWjojly)DtC5Y z7cW~6#QsgV)uibMdQf6d`oojZ(Y4&h%Qd`V~0hO~%-gJ(J?mMB<2xB+wnv z=qb(gSC=q0`JV#q!pPRwF`ulK8m}BVWi0fqrJ693$n#bTzZ26Gw%#~$bBN|Wp#RiA zIDx-W?`5gbh)t!HAU?5f`hPTv#Rh=JCVxV7azA`-Et>>X_e^;2sPar~Sg-n%O!m%a zPgjq7r2cY^)6Kpu{aWja)ucjXG!~>KmnS2t%^?^nuQE60pdTo4d+L{-S*2D@r;kzQ zyocRJhlm%_9k3S&kKb-q}SEyW!v``9RXgyBe#JFQ1FurI#bnt%KqtmYuq7&e+S zqd@FDPGgfb{746P@~t!ES&`A{weGir5D6H18O)x$*X=$syKR0^(m<4Bb0oDKz3RkW z`qgN1*Gb7!2yM*jHt_Mz3{$IhM$CmvjT@DnK&Zn*t+a=1!K9L%5~&6`b1S{cG|yi0 zjGGi8uLGx)8Dhga(0}IeWbO{-neo%V58&d_xkrEbb8|;yUQBuY;d{&#Tca=Wrv)8? z?>`(00!#us6RSDuJ|1x4*zsTu%~B4{)^J8SF9)zE84@Pm=@>KZeQszm@j)7o>s_JR zLuH#}$}yIXy6AGc;SF)Af$w%onlY-3!=cd9K(jyT)j;(oP8Gg`;gRFv5)_f|obz+*i4-_MmLp{H#kJKjEX0&{k>e zK1=?Mpr?9;HL5mP)qsXrc_c_}w*xwKMD#t5m8TZUZ1if(k%gl6Ztnv5ZmOpf(+QjK zLDs#PI?=P_7=lNO9daQvB3b!Qs$tWS!~#%L+&S-|;eXY&NIoAb+b8>YTheFft`=)n zNr%A~wcTjM>rdJ0asn`|js$M7x;76d+1FP=yH8h`KEM_IZ-L0_W9IZ(%oZ;ArU7>-oj!NC0@vaYNKRq&K3*6VVr=<_OC+UMae zI$w0%8Wg%=Z!!jE9G|$#Kc#0ad`22sdHwQ*d83Yjv3USnixA=at^0ufIfY!`oCK)k z9((f_vUk|D?v8l{-w_8m^`C?sSnAVjy|S(I2aHpq!_&BtZ10pWB<0 zCx7f!dzwZf9K~1%X9P~_NGpZ-XH5D<-L6IgH=hV^fH;*XFj6t|BTO7a@t@E4+uX6@ zaX0=}|J-U5%X}h+?fuQ>&~7`!igNwu@eJDNp@HbQ|QV6Nn9FMJK`gk(r~o1b=1n zluA?0VQZFXHR6@%?A(<7RxuUBvRYttB>VD$l zju%Z3mBP|?{9(;^UXh?f7(Jrk)0+?JJ4-JfW4tr>t8Lj<c9Z`9X|Z#&k{ZbDS!Fbz2i`7UuBJSAC?!#8?hRD=XRA^b2+Vr6qj&p zN;0gXc6G`Vrmc?h^W!PYPH~b*LKTY|4(t_u;s>GwZG(@QO6%q9Icl~fsVmK=C}ODA z>84=2G=NTTX6RXw=hmx1v|r0PFdIH8|JQf?{#Z?Vsbw9^oo}0Gj&PbAPZ8Zb%%M zNu-vfH&>5eO%iYTa4RhxoG*Pgy6pQnpcWWKIC9!|gmbZ~&9lG-V|BWNdTeScB9YsQapJf>{pik7G<&Y;#PhmGDWxYorae<(>ScJThQ;UEgYXr-rSuPd(`V{ z$bWdwvxx>ugA@g|+3YL6y;ne`u=Y;xOvqDxUFia5^p)YFY2&=PzJEr-HH9eYArbC6 zGVw~AMa0-Q`6zAI;b~3xo}rzoYg_2r?nXkqeivZzDEKBl9gkYMLmfSN_GVt78`F>Z z1BqeF)60?!*xF}aL8g)Lz5da+tWwm@iFd~gYES36u{ZW($E8NPOUg4JDCCe^~VsS%` zl4(>KN6Te$^QFN1vJ2t$Htz?OXaln1bfR=Q^*EWQCn0aHy5?pHh`UuIX7G6ud$WRe zk95X|8L^R!U6RweDQe9=DI4#3Ei}{d&aLLFjwfXkzH-5=3+kv{TVre;eUKWJJJ#+d zHAJt0VU`7t$bSw7xZ)osQsQTnG!pf*kZHtx^}l%N>f}|J8-6Q6e(1r&%`);a7n|IP zhEW&65zJkG@{Zxno5COYXxa03iSe%=WyTLRAL)wce|SkpivJxr^`lYoD`;g!UAzy6 zOX`aKcqVcy)a3L&L3w?Rz_%G5bC?jLQ&Dv1+j?p8zc{l8Bdx3;(Y(1cA^z+PmZggrCqqQ)*R{U*%jS(yCgl3tV zg_A|7oHU=CcWomurew${31NE@2H}+`==%7hDS!GNE%=--rpkq7U?b3}O#0!v6Vmr#6gwoK zv3ISPmM2rZutyK_8UbNTF5Ag1{OGUvFfZC_ zX(W>ThI|mpxy5vRaH_NKQen;WDA_VL*!ys(zT<{eG$eALn<>luP^EHs1|)4fq+1l6yfg8Y75slEgdpTxT`(Iyi?q+vSq+E}I?n+jLC%@0P9PWK<3Wjh^_d*v%M7mmi-opP5ypJ}^o4Qv_`y$@XQ`D}Oz& z0X0K@H+j=76AoXqGj2A?A{ajiviK=%`(Yf}^_A&RHp8YQ9vHmVWVDpT;C-ZJsGL5N z9C0uf+1neLvH;Y|(BR^8YJTudcW9@=wZnvT8n1RQ%j%2 zT_nXH_X3T)DW1hN@B3`fa^bcQ{*m@oe~=Lg#^_;)^+Y*ncNX)BSU~<|1b_BH;JtAE z6Y|OOL<2dJiu}U1(ZO^4hq+jAUm1;599`*f{KwBC-|1^^u5^NboSlpI(e^PqX4olc zOeWxcnz>{w8B2bZ@NmqDlqq{F4@ZnVSw2Y9q1`0-UEr3$M8#b7z{xgTiK->g%3#~Q z#H`^h$;e#pM_Sxkt*XV(X@5|M6<)>fLk~QP3GVT0>xPbksdA4UAo}SxOeOW_2(u@< zj9vTN0Z9$PJqMn$%LTS~J~65rKyHqjTM z?MFj1WUZtnSqm!8!rT#VuwOmvwtbcub>2=b>)co80X3`cA&deA@_(9C*{wy&?!Vgj zUgD_xXhMs`sBK60qB9n3Fla4`zNXn!p%&mm=Mva|zjd37i?)ygwyD3`sqb(hD>jv3 z{>6^;NMt@uE%IVddWs@bng%2MF~Q{!STQkuZ>FYUGG5kq1}n>TlKEh8P?=)})`h2}fs733^o z$+c2f^*3Ww(S_Y{ce_KEQs>leUG<9_CKaOwU2EgwwBC6)`wPTr)ua(ni7c!Wtb-7=Icj2`5FW$Dm%Q;}_K@ zjG!fk%Go16&c}^V5koPhs!n5wS3%IMAjz^~cH{Pn`{>2QNBW2Le%!*Be0|N1n(doD z>^-zfFVd8rwoJX6bjWRZC;ngQL^=t zKdDX87wo2H<$n`coU`Pmr!@M-WP^uP4p#$rd%M4KKl^SKlS0eMp|weMe?qy5KJ#9P zWBajSVgN_~(2|~GGbJ!QoNQv;^D<^2>|!bNg|dK{%1LOHs9EXaL=cYom{n;$9h<+D z!I|U|s6)kO1b@c!UQzL>V8PAR9rlVoQ@hVw$z25w)qf{Vz=#g$X)+PzbKi}lfQO+v z4(~oa43`Qe1u)z*z5*&92xiVsIHP?^I*isLmn~7%?+P-@evFibn{7Cg3mC_rN5d?} z8f|&WgOpxu&t}$E)R>_YqSN~;YDHxVy_|~YA$e=9gh)q5)R;1ylT^%*PU%!fZ=_RO zCFOKkI9h4;oSk!ZzML=T(|v#VfB0Y5ef>ZE@89>iA83!%rDegkk||_Q*A+dgiU+!n*Q`UgRCM zZ&A^fS3#G?)#mign>w!S6Su6}vF2$^NX?4!Uv&A;KR(7^m>eRDX)HJW5)}T`N!_m1 zwyzvvo?D(i-d4|y`}IirMy_w~JmpU};wbotvCC?gXLhd3LYuNB1#9Crc|Yrq$@p~k z`7*_gcCq|MpX5qPQj2}}&7&>(FXWo@(9{<_dv(nPIoo`S9wzj6<<`lf@{^a#eRN+i zU^I6`Xe<-SR4xQV@zFc}!#%=`yqwkUASX|;+H=w$-RwN|YC1ykJ$FdkgejlsF4Yu# z*#Gv-8Ql*9XPR|{`I)&@ihA-LQgy0-bh@7_Z!KEgQsr%|kW6k0P#z;%B=X{zffIdU zcjLgZfcA*sUW_L@=&z#Igj$JJCSwu7sDxqyOfg4i2>q7V^B|rjKL3kD)=0joEVWq^ z1|*YEB0{hbkQ3$QK6)R`JsNdaiINT;wn+hk0|UQ%x1Yo7bsDtDM00^;xYKk{atRN* z0T(>*jyc0$g9t4PX;g7$jZAjp#0h+llV{pU(keB@HWalqJ(o&$F6MXRn%Z+$Q+M?bknN8#Z= z4(tXX(gS7pM7E&-FVQ<$X7Iv?rbkvdJHEF>RitpVEaM!$lj4R0+)7eF6`nNG$ zPKpegr`A1D0an5-0JZC*^kl>ZHIh96&c!rUE%T>fQ3|CQGvq) z&DQ9+|Aqj>4jvsv4f_o2EnJ-G0jUoF&$EiLaWuPWurZgaUW5ak4=3B(ek_$Bm$UpJ zloN5z(jSjf?`Q! zMzz}c4wg_YlO6`awrD00HEjt5;7pd7R8-6W{II*DwdBudC*tKKs$u(nEH}czf)8#r zxTg9xE3*$^5oRw->on&2hBS~M-|m5GUydlvk$k{ku~~|Z+USX2uvl#MX>6qdYj(4u z<6bqEYU!=mYxC>vq!h$vjqjl-MMK$%wlMsNvh4H&rP(^?L7XlDi?3yyiAzYr3a?cI zZPGs1);^?H6JvFi{^&Mdl6qBNS*bjmdMOokD+}~s>7OQ>JDB1Oh1LrX+RXpZyx#P9 zTyH$ZcHvsAMlMO0{+M!HT9iHm`0(E1lUwH;;QA>5hbkHdabpvndr%pE%28DYpm00_ z!1=R=Ydh1>Q$R$JVt_t(crlZI!^gL-UANw(ryi+AmMv)5u%1J4mLx8{LlBC#CMG2a zqvHPFjL9T;F}_8#p&{X*uMt$HDPsA?Q8e0Zq?ZeK5& z3x;RSCeTCUw+2Kd2_cVw1x%R1h8gfTN#Qdh7!u~mq_I31G!{%_ctIX3glnLX*nL9C zC}q+Egwe54qS?^S-BC#~iNYvopD-aYHeLj=X$%@Xo3LmR^vxjqo4>}x)?ksj(UKhn zj9AWQ^4VVVy%_;s-W&m6z!ks(wjZC(;&c7|1Ole_BFN`|B^&=`u^|SN1sNY4JP@a@ zP@*c*_NYc@`*{3)(9b#P9Ypa01?g+#%sR?axY5}`0VE8)xqV2K0H#cP!^BUye6ObR z&#WfzNqN~LwvD;da<(}q3hf^jAFm1(bzXog&lFX09AEFBd9PH~boKf9IeU_W+=mCx oPkr@Oyso%p#qTGtonGumCEWYxGKCwdpuGjlo8ay)SQbe53s4v3y#N3J delta 18009 zcmV({K+?a9u>^pz1dtX13XvCV1X3Uff{}+g0j`n37JudQ?frTg7T)9Pc}@IxO>dML zjO6`a@1Iy;9ws!qd3buZB7?;$4E-|Y^S?G~8KvIe%P{q1*DJ%STH6vfk;= ztnPJVHh*?Q4p#?BASiwR3|^A85?#esxc-gCG|HJ$258vnIcywRNi@JeSCgAc9}ju+ zuBcCt!i&VQU*s{J``FM%b6{Ila;ti!nG7eLt&7g6ZOs-*Q$I?g`MZ3NqWFHOjq3E! zRJ7@n>P@FM%P8>Rt(o|%+DW61EcZs^J@_@no`0#2EpVuRg+Yz*$T;)U`0T{rI&Jh; z*R(~k>fe?Ld~~vweNU8U+4Sx{57N9W!@$$spFL>I&r+0r2vTR#H;k)~wW7%{qB;ym z#&6Y0*Wv=E&?(swV^m~*o+mKL{i0;%Pg1ZEZYyQLAYi8}b_>4*U!CFm2S$1Xd80?3xqJ`JSEpLv9TnS!e{L9FA7E;)G=Q2)#xC5JL~-hu#m%Rh$QJ zfm~byaxnvPaS_PH3COEUp^TLTGfbciUy7+`PzK5fWtg8s87&I@{jfPI!zL@(lVQ>h zn_6wf=}dPDxj0NdexmW`$D2#qKHRRq;dzrh0~vpV7Wl5nau8@UvIXWB2mCz7}Nc$AIl znQ6ykQ@_&e@!^(s046chC_B8MJK(a_2F5Rg+W<$WV1%T|ygMhhANG6MkMS#*dlv{p zqr-ol7%&KIxG98*mV6ABC^!K1g?Ux8)m^JQL9YudaHpeN8!1F^%J;+@O>mE|FhR*= zH(B^1SU0j8bA9T+jT_8%hmwNOYwdAEEkQJ6Xx%ur2@eiE=(?f>d2Ce&Zd2R0 zwt4b<5JUtjSSPHdbTwQ*1AK~?7*XgFDv~!6LnO|2QuJz5vw-o&7v5mUcZ?Gbg|YMA z1ottP#_*bZqNhb=h}#AXM`Vb~nZ(31OT;0~-*`ZpdOiX?h2R9?O_JE7sCWZsfP{aq z=j3UcatZ-kUbvj7AgS6w7rMFwCJ+<`$;3`Dnqh9Woic)Ca4P;F*Rj;87lbYiz@!ze zE`^yB$6_yu?|Y zfMpw2$Q$v~2r_#^pI|~wJMl>et2lBw;it%tU-k`v_sGIpXIecl$K2M04GID!q?6JG zB~p>c%IsYc&UAdn2jX{#{Y(`5Wys2pubBH~ z9)!r2`xz|r13fHP8DwDi4Z-zSS?t|f%F~7CkHhIslSu|74UCrql+<$*FCj|<8K{gud>aw()@q}xtSur?C z>lrwE*>T+(Ixi_)9xXx+hMYGF3>wyu$jl?X3QPdrE*Of}{!GqR^gR0>q=E9K&jPKF zAp>2FMQ!w8&LmVrC*Dhc)?uejsjbj^I4Xbq|3!WTn=*=DNixB7l8l(fcS%wz!hrld zwlDAz`w>kDXf$JC)|0FWNPuA1vWwPOFv`CM8rIDO{ap1~;p zX=^TGqC9uHX#*F5*g=p{|0^kLpFPwig zGZ3i}ZoSR^uu3@|+1ODjHxP}JL2OeAZ-$QVs2a_?ocr1T2P{SLDWl+a8pJqV;Bgl# zXVhVeQ4T}F42JPLIWmnB2Pa!v$K?bc@Dcl=ZT!$NYa%telH4zHTMa!Clu<{j^f>*; zCA|xbc>jYw>b-0vT?UA|P9J9tBG-Q?G^LimJmp45b?FiHp5Y5*)Bkfe+eW7_h)o}w z8tX~`@wdY=^(+qupA5T08|^dfiD<<#wHi~UV`?OJjG}w}kp!&?n!s!?-#V~FqgYq8 z)qEV-!E1Jqf+{-a9B# zLI@B_5|T&Wq)1UwdgqJf+N7}*EkO-^*SP&!yP}bGe1cLw|kgy;KBt*u^X#&T(!v6M?ahk!<7&w0t zA@v^!Wi$+e#raeqSR7Cni2!Ila0P&c0bnsHu(%Wm1P}s&B>x5?(NX{vhzHyrpeqQ_ zLLy)oGEQYA${P)Lbi(5H`FjcAhVlTwl9J;5ztRDUZZI?)3PAvLAy_Av8*WD^#1&wS zgu-E1?|+owmT|&jQBpvlr>CbN#0`HVh(tTe@$dsY;aDet5ex%Ed%)}gKSu`WLEK<} zMiV6C1eiF%F~0-GNC&JZ1Pue=9IkLE41vL=JV4mP&;Z=-0Ao!ZfIbR__zl+i4d4g- zIUE335d5!lf3klTf+Kza4&!*2uD9K0QB?ouay;!VfILb ztM|X*zpfalXKZSwY|i&*hGRC{*ZwCk3x6=|1745#8DRp;Qlw$?}J1^P}~>z|IGJaDgQs3 z|E%)AGX1|hQhVU)`YX=;8}R?cA#QM2?>`6}bsu1HGSEfh#DMs3s5$Jnc6DL)@CR=H z9o57_aB@&YIN~TR02Y4{1d06i!!c@bFPOan91C^&O`G3-)1QKKg(F}FNDTbvBL6ht7gxD)_xW`BSK5>58=q9s58putbqFAIMRmIMGjf0_T-1^NJu z#X4SBN<}75fz=@yqV2RJ3n77_LTeIix9W(^ zQi5@I*0lC>wW%(!nzeJAo|5}f&iv%A?@A3@t+&X50GrynBb!~^#O1-2u|(fP=GzmL zjd)MZQcVK$q8{Vcb{V=IHk*G~P0!j^saiv@q4l%s0B9jBF@Eh~2Jq;_#9 z+a!k=@E4zB7HZTitx_5~uV5|8ekLcJji8}dF7Xg!)|n31cr*X<6N`&hc&N7b^jPQc z;@Jc3R87nGJt5rux(TuuWh-C&>o{M>dN_pPH~&?cLcOBON0-I3>F*Gn-^1aWJ$t|% zI%Y0pDUpBX)M6r6JS!Q}9B4r;EKK16Pb)XQ4y^m!d$s89fv=CqMgDGUs}k3=dqp%N zmL|q$v1ZmUqwwj;HW>J}7MflaqwM49wshcE;k9I?)9Z#T@_*^m;wJiaagzFR3{ncf zmh)`0Eh;3IUa&dx_RE>&u84isGWRbSOQG-wtM7G_R`eApo%{uE%y2)ZctmU17rc@1B&J1iajz*inrF0%Y3%E@mQvXu z5jS_Jkdo7PT_SC6epg~PD_l6~c{Fgfe#iVa+gAqsmHO<= zH;a&+4dgvFjD9>%NKzr^w!P3!lV^XX zSbCkf4Q$uil%@3?;%D3H6kB5Vkf+1*Zt#CL)RGJ2obOqHdhEZQs0FdvLd}W1h6OfTc zHRmd^;$tzJfU4fh&CJ$dH#?jSvae^a50thS$uBS|qMQHx4VRbjf9buHd}YTP_ftfrpH<$7v?-4Zyx1f^IOq@1<{~&d3&5M0-0~zeTp>a)0sS$>F!U1v&Yk8`CSbCV77!`;gy^ zJTpd@q{zZqGht2JWWXKL$mVqx<_>x~@6)f1*Zo8C!m(@97Wu)=x5Wc@+_52x4k!NY zG;702Je6S*gvLgr7RlbXXk@Z5wL;Fx`qnK=!Oc{h?bn`|W;cf}`pS*P0`tl$o+-N_ z7gj+-oMrw`Q8TyfKl?h-iO7FN(L`Of_EahUt^hOKv%lU&unB?N-yODIRA{S{CqZ3< z3rjOs2Pt(oDYC#$Pg#5S_OB5P-+rkL-B`0BBioZZdwXVr9F{HQWxmn)cHqRjFmUP) z>IN}Ky6F%zq&ToTFk946=$^0?eRcy=pUhqLVPEjkCDcrlp^`w1+XjEFd;2I78)_dI z7(Dk}Z!1j53JQlT9f<5+%5Jl~L#My;@Sdq~vgYm3(4MODyTAAGZlvdpI!4{C0%3mM zx~?b>QGdbbMUut0Dsyj(L0W35ufy(cWcbGsh&>bX*A51KvoWir8+sLUb!HG)BGw)- z5DD6GQA&vDxp4g^=b3*nZ?P%!N#ky8We=g32hHmn{(&oM0vwsQGwo>Ex$AsfR1A&m zu+lZX*EgIC8@sYy$XE0j4n$s#5{n$C?=95KUOFrCX4*QAgd0AuswU+w{wzHZRB1?I z$`+!GvRi+rAbr%Sj3)`(;8L^XjwZ`L3q6iIE|wpnk457}eNBIBX|~xb54}n~;-F7r zKK}uKxiq)*!7*s=Fh14=m{87DA-aFi>Xg;n~>8k0{`(41a5S2AOg`nV}rp)O1iD2XrKJFIzS^Szo+5=U`s zpy@}THa>ytq=T53`(hposW(}*r+?KW4GmxoFXzdOU9wAJ#o65|Z=~0Udgttk|(gyvb&`OJR~I;2$v(xgs0X!6ckIxxuM z>(bB(kMeM~{CHMQmFWJwyk4K1sO`SMJ@yy|)|@S^PBIN1X0+i!Tpf#{=t03l6M8!x zf1uous0+;(`a@Ym&wTw0aXQhnh&xPD$&>u~rZ#js2J-QBz5TcJl*`wQbq~hj+v60| za>>`Jd)|K^Myy_2{UOXd>3qk$Gh1a-Vm+E;XVh;ehrK~9YN3BHs_}wxV_GLs+U6md zy7Wbh}C?$WdlAb7!B|QkiXh` z#%#<_BoBTfA522NJE1t%z2NQ2{b*U|O4EDnt2}=;f^E|{nr~m$N!uA-xG9YA(w)Yc z&`QS5s5apjnd>_WOs7nA#86ghY@pWYuR}D5S#Oiv_bEJoLj@ceWerhvHY4)Q21o9K z2ZK$jlFf9^M?}5Nwf%eoM|PJ}zxz$sW3LuEE7@;^n#q>?ObJi7n@ng0w5(ZHx7J;Z z3)_F5d*of95a?jjYr1CU&od7MKi{}3fL%{d+o*h5E##9u{v$r52rryW##jssu5<00 zvwdZku@;4w+l;5VMK{c1!;^h4eRs?TMx}b)7)w!P3KjkGPI*Zo`Vq2Ql&aTDN9A3c zVp>%F!}D&vVwMj^BmkZ0uUQT~UTZ=s+$n#L1`&DH(-T=U@Y&es?}wkoU?@7LT4IAg zod}}mNh%3gcU3K)2@sT>H+h|RHG8z zK44;gk}FW1DQ9E|{6Gw3g|Q7HOV~^H%VyKiVKHhy|%V1tFVnxdY7oPL zB5-~g*xs|@*@+G1gO|?3%L}!hkL5I+jR9Vp7IhDN$E@FN&UOy<&nLV82CJdGd3C%0 zm^B?prn0mOc-2qp3x{QOp)yGt z-?w+njm%T`8ZS^B0InZYkP^yT4%O!gY8tU_csvWY<_m?1%CHMYtME)$4iKl83;Vfp zx>TU##%G%$fmi-60XoA9pTCy9p%_q??XrI1^l_8Nr6RKlPAXwH8h@mKvA%WWrG6HG&~+Jrb6h z(e}4c@?YMMl~-grj4Kb_B(8ck(PU0^fC?Mibs^3V9_mxfW254pveJKDVVI<*B&DJ~ zn2oazdO19d9-x@Ayt7XbgGi z>j+ySt#^qRF3~MbgX*eIV_fU)+Q0t0>n?vSyq3GXbYj#Pf&gHBGvU< z^t{u3bz8eD2M|_?!i)wE?)jZQYg6GH748|eWBc37yUIkjUtD^$MK|s{Ur^)4%BSU? z%nm5Bj5e%^3YKD|TCIASq&0-fLtj_RySCLNAE~TUwim;r}lCeL&%bAy;&*isQf)Y zzBXu1{&;xS6R6LQg?d*!y{#4|?q|(ngxuY}UC8GMek{_iKRJ;0 z-QPE95_x|_71o~n=1GZO{`MHvB5|wB)?%0aqF0Beld5cQa7jdwRe}{7V3N@8Ptp8M z^U>+oOrbf3@;eko;e_cIo86YAm8oRo3D5`U-c6=PK`x5~akw{}F!b1F7 zEDXl-%wBhMZ}`j)Z@T$WWM*%QuyB@? z)Dx+d9&e{JTsxZkTFrQcKv6%JyF3i@)}e}%d-fDHq%xy%TZoG&O#l@|uDu;l8ELn? zC(JJUKy}w}BPo@HdrL(|y2iP(tH0r7VM=rLSO`PQl3%}Ys#xn7rk?mt&$u@V+9Nb} z);xcSrV$LATNWQzer&C*l(!VUUP&uEhmISzP1&K%a&>>nEo3#_0@bj5 zkuMoRfm05DkXy&5?Odn5No{KoV0_6Yt37uquX?!@Wvp+j$m`)W+ zL!Ox1S=K45xxW!vqSb5r{cVy3kCE5*e8NT?L6;Y)=&<0lQ|TR&W)%&F&D<$aA3lGy zE})}{POIqJa#-VbP?3~7Aze?OlehL$sBi3<0MTvd z#cUPZid}Y_ExiS`rz@^yoUT55pbSU==5wNofwG>6_Wk;~oi%yYJ3ovNBZtYiHZCR_-T>zT<&SyCuAa~_2++28u2idSCQcf97x%=D$G1D{FeOcdqUZ#`3?0C zz0@6vM0gDuCLHUUx(7?%(RdjWbRs}q7Qf1dwmZ49wqlGuMEFGUSc)4%5-$C$@ceaR z8N1`;tlXw6WXrkq5qzxDvJ{zw>N5`?`bbcT8>3NNVbvUwF8eF*6qZTqiqn6hA-aJP zZ&Gg2-s!J~`JJZi*;RaEO((yG)d`mvt^%d@NBF%T-E7ro2fdtd~tp#|u-US5Mj( zo{H&{ggYxuFH6XW%7=VjRUvIZ4rYTor8ZVvc_jUv>938jWyjuj0J`4;=r8o7uzycbQH`%Cfm*^s_4$=pk8P z6=fS2Sk1&=)#^u!@Y977V@_r+-A3=&g1gFvNu+ki%ZN(xRms;gFMYopc1M> zMHsSIYeGn!`^_hykD3}k>yk<5gfUEftmvgW-@W#MT&0+*g{Zbo*+vpxuE8J^@fQ*S z55wc`q>U{0w7Dj~kzl)7&sN?2y3tANw4t~ok2#;)Irh?Sm6v}`)(;!Q#19=w1krSZ zTpPA4Yu{mH%$hy%%6h>gNrz9_c-fw)F$C|V3Y2!v)ErYk^g)US_Z7{)sIf02pi+I9 zc-qg^N&Dy{aZgF%(LZV;=b~BB+1w#o?9U^_lv|<0FekhP<)?M+t3pTAOd68up+GWxO-tL^gbAVN=VuuAaFeoDr$=Jw>CGLTL@p`wyHQ;&>G6PN$LEE0I`(og zk>U8-S2HOTPKC$se>2vTl6YD`_m?B#M)3lPsr)?8%9wwV9p~`rG%}_kFl}8fN{nK; zt2clrjZ{B@?8UQQ3s2E^z}o{$elF>y+%N_XApN}7i&k~=#>GU|-u;doW{~ba@{SW~ z*kpsVYiIR}fKeW|ToLck3F%1bmp&oU2&b0y?PfnsLD2^twuS};hZB!@Y=1mUH8kw4 z-{3yGdMJOAC`zAPV_gzFAM%J+ccB6q!9pUc&TA|G&PYwSlwz?n@M3(>aryBks@-9F z0Utqdf!8Y91E#buD{PAYs@=6kW2r;t-RMqQtJGf4(W&qkLqxyW2D|yzBH2*fiR#Vz z9?hn0u_OLdO_L6e@YstX@64(2Z;4klQc&=wwM&09s8fy=y(`zu!c(=*IjN>MRQiP* z9C!MnNZ3`RO7Wtp@j^Ly@UPNB=$A8fy%Nv+dT-n4q@Kt#zp zX%<0|mm5g5@G=V$dWf>*U>9dx{zw+>;8xb|5^4}o6i z-5xwB_sJ`FItrdTvJB!X?&b<_zg!zTa~CX+?muE_oBL1{eyeM!Xsg#X$;?Z5-z59_ zwF6EaJ0C`3J7b^GS{bBN+_>>;L_o+bxTk-gfuQdAY4lBHM##%du6OXd$&u=CkE?pV zzFVBr0+un$IRa^XxlJ^JNCoDZQYU4YYV=30z?&2+n^_WWJq`$nyzWx;dsT%YYX>Q$ z>^kDfF-DTaY693vMg%b#33^_*sHShUJFYtd?5i85qx(`nU=S9UOV~D#J-10$S-gKq zZ|OF@8SHnH`-jWJr}U#^haY@)+N^{6EsoC;$*j@`>erQLnj0y~mFlKLjwibaq+7-f zdLHPnex@T(qPBe(ZDLtB9bw+zTVN|G{j9JGNI;u0SNXZy>>TB4w~g zzD~RTVE?UIq$RELbghot2(MPj1I?%_cV4g|TX-*{i5`B@tll(`CTL+j8Jd3-Qt=x% z`gr!^QC{p)oVaCrL>=8~oU+U0cYH)e2ifYzY$25Ru#!lX`k8P85wD9FalFpkCseTc zZSKJ#D&<1VKwqNroms)-v;OvvM#S!pj6p*Cm8#=;g5mZSrv%qTY+E3=i*BXXwK$@W zcf0!s6UtkXuHZ+p6>8s$?TUZO_uc|~T>TJM8kN8@VeGR_*Q{I#?kA=Oo|+2Otm%;q zC}W-$O9+`?LAj}q*X_DH0QI3J&vmlrSC z*_{{RW+%U#=x^^GFKbv)gBfQ{# ze+^q0+{X6T5H=nze1>qCs|Qq7@oxwU!utL?g$`^QEt-woysFax5v2Lb?p{r>&-1Vt|!1l-BzKjFXk%V(@-sGzOF^=IOL z`{d;j-T*%y5P*kY2nYayL_h!$)JwoWtLT7Xf4czwCsh@0ivWOtf4W6o)4w|Q{7V4# zzvAEk{Ii%A0>xS=fc-xtHwOvdS}N?#C1)c3<3#mhe~f3Q1B zY{=hh{AEK?a{6ZiwJ906tzg0*Oii zpk@vL*dpBUeyv$l2*79f3;Jz=L;-xL{UE`=k$)Tn@o;lPb@H1Zs4o6x|5iH`>J5e9 z%}ya85}^*Yp>1c?ax`8%Yon6be|za?Cj3031s1r-KBWoE^b!^<=I}4vV}vZ zmb~lP%dgW%eyer#^*(|pJoL)n?dYu&r*2QsCzJh7X&I-8n$TVwr^-JRFLvug)1iyG z|L(4gja{EU3z0utZO=@y7b|IrvWQw(n^woW>XK{?a#}H?ofSP~v-N<%f11!SC0p)_ zbHcN4vF1@}TA7UMk{#!GNnTa=m-+_>7KnT4x?+OWJVVy@nZuF9N)iLKih?**Mhkg_ zSGmQW+nEpitW`+lA90vXa3q%`u4VNZZ({{UR1G?>$E|zhOLPZ0Y?obS%Y6y0rxPY} z$+prRe!upkuI4CsLG76Se|C@(Gx$zhLZ!gI?8}Ec1&5E?fV5Rkz2$@e!xdU7{R=oq z{5bROG`+7Oe>}sR>A@oEJu?*=;|* zF|eVKsSn_!jFP`{{8Tc)A{{iYFa!x3=82Tc)6|@4DGail;G=&D*yeTu!c-FA!r`Y7~3qa4$ehY0S)qqGM>!~izTQX`^jGOf_PujC9KFef&48KcSfw`d4 z-JvgyQQ^eAY&}*<2+@cUf@iy(x8@v@2g?tl{mTRn6|uwxSBx&Jd%T$EU2}AD-64(L zo0~Zi#knj>egNWDzDrIP(-{vh7Y0&qJ@t0Z`Y zrXohU&YwAlzQ`c=Y3`&Rs+x)>Vuz7M>q3 zX&2Ehf29fQnzLAkPTsb)WSfm7Ut($$pJCRUj(m8>&`G4h+d-(b47JQtcsbw(IZ|bu zV-j+43TGy4TkVQqGQR2|kJ!FGp6;v6GvXwn&bHur{K)=E;`w!?W{=GnYqggx;@N7$ z<2p<4jJMi(4WG_b5Jt(%EU;-&U#pj-lKtUrf8iwhOBDV4e(Vloyy|+Uu|GSn$e29B zf21^be|u4s2G<+?JoO^7R|ERid%N@VM(z!6=MH&=KY7k?4zMz&@JZ0Om2vU&8`-Qf;KQDE++M_g#61OAbhxKy*He+?Zs zrHVOTO+TYFQOiDW`KpTGz0GWYCPv!HfQbTk6dyB50zLh{BV{}On3%HAXfF$m+}gXP zkgpM)fg;5%AC@K?Yg`D|WM40rM<8$SaZSb-$48Yv?K~p%1x0z${5O&4rgd+ z|5-R>;Hui8z|2`AN+@#ec1Azce@xv-E=(C_R3O_Fpn2sCl-#D<(gvtQ*Vu|-6#+4O z)M|7BI@+n)FXaUgWI6A*jICX7xr7M4zhL+nlFzzJ%aJZ>^I}-PZHkg-u+b#zX;)XD z#(d3CuQzpWo_C+;L~NTpOM)(}6un;;kI1VzBvOOLp-OVu7IVTGNzI);e=R|~FCU^K zHCUuFJSiv#{7M`pp!zt&kOcGdnaqyLf(?Zy<{bMwTI)%-h@jewdW$%cXtA>|;E!i_E9!(^UxJUxs5sjYGOn-F=W0U_W&@PN@eI4`3YN_$ckyFM(-&(2(BZ)k3 zrSLm3O=0VeBR7X=-UIqi{eu(u8}(k63XRxQS_$G4>!wemSZn}je{Aw6L?`#d=hm`G zKy}ZA_l_#h#D?{%Kgne8eD-wpxJT+Q*Erqm+tRPKu2@YfL`GvlT5@?ZqS_pSq4Fwo zV-EU(61S&*>6uk()pYt8WzKupZFGouA>DENyst23v{2`Z_1IF}k+P4Cl1CVh1i90C zL<9TcyRFIJ!D2bJcZE4tZoAz@60f@T4%&uxYW2&*$IR?Jk&~i z$QDd0*(s4~kTbW^n@scUCC|7?5%M~4N|_-xtOIQhPv-7Wf1VjX{rdne9-Vvimp?am zH0H&W*B`#eT(LF!5`S9IA^85op&-B{urslmqweDY7mghd*3c~F&}8$)86NX784(&@wna8Oh?ryJf7mm2tPr=%IPidN%stDEFv6in{I zDq<$)I!~;Ne^C8#ebr!;ij|uL|mmd($)VVmOO7FUwAdjRG9!|e|D+l=9Z4(zHN~Ct9vWU< zi{$g6f3khDkGCa#hVE*yW|edpd{NttM!f!%tu7}3)9Ogz2CHlHaFTs}6;wVoPye)g z(|1zOLC)*UqY^C2gwz_Wc745vJx*90Qowv4{cFlI(&zfx*&nsZ^0s83S9HcSw9>5l z5YzdAc9r|ehjCJSo^eE5D60nt1*q|!eUzmee-!j3TATxQy!fUlz>MKY#T^XXFCy#8 zYET7FiDJDjr;0wWqNRNv?xOQW*R4UJ8}=q+V8-!@tNc@X*1~6`k(JjkUzj)Q7#N!e zu(b#gzTdhJ=$})__036uO75{Ye<6E^P3!KMSMVKifKxBV5PCKvL+Twmp0=XY@S@>< ze}!2mde2JI%abNGY1SpiDEW|syOjEn%<1T3s)vICfkKMGYtD*>scrU?o9cleO%BKR z`!l#4Iw!a+JbX^X{>W|Yrt}*~`z1wS^0EQGrTx1_`x^9vILE`Zh|ac!=X7r)J}5NQ zI025Az%VRUM>b&lUV1~?4t~O3wWnz$f5K6Wb#O-Dq>i*wh=0bUU)1etByjVI@CJxe zi2@@PGe5$_F%s=^j=Rm8g%3#!uTVk;um!<4;rd1#R^+6@H z{1cK{0h%ki8dGJ&E357&F79~I1W_q0ZO0$heCHJjI)u?93O>F0kiN6@;xWcMbHCb_ zZABgz?GKXi<1JxfUji~ds?**M=&~B(7px8pkl*3MZ~iRdQ;?E>-8&AYfA&?@NcUlR zal8?$v3G7)sWq3=T1asT$EGC1Dr#4!JYm}EC_g`*vg{Nmi6m69sNuj~(ICqk39$b;D*G3f0;yTNqTej_|+uwh7Y&W(!u%CXQRu$j{|CfQG_F>eMdMK ztJ*vZTrgIrJIIGHWCHU{?dVO|T^h>Q`B<( zu0ZvltHGPS*9+s*?Z}F#SIDg7#G*FF`+UeyKF#3dim3-fUafOKf6wutALvB5Y_MpJ z>7VBPv}P9`>yMO~g;6!6OD}HaHzHGXE8l5&i8GDE z`nm=E&f3E9$>h!5S+PgGu7>=F=RBKepfpHPP@B!Z;@f)#R0?bF^v;Al)z_6SU`AgV zE}AyZo9k;NTvLdWe;yLyt|Jq#v{^)qeUp#Ub{(G9bnhA3nYy-xuI+9l#Ors_rq(a+ z8;_7m+J4@EX}1mUN*9`_$7}5IGxkY~RWV29(20N9uaNi_cqap3mFS44Bt)QG-Gxp| zDevHvMeMH_JbfWq=273FK1Z`hwZkZNg)S<}#VbH@-P&R?e?!c>E%@=2!YZ3;{e^;W z($n#%l{?hYlV@+{1-dc)m_LvhwmiKo$$+hW<`rZb3E%4EO9j19Zxh&uh`8FWm3v{^+*VSZhFQ zR=S`-1HE!re{GKd`y>`ODUT^b$P>D7mD^4d$ms5|Ed3qA^ z=BjIMmVmfhHDU&zC$TpxX!l5Ge3%g%$=D@1otvW8?31$bp4UP%9q-&~zUp{VHsLE5 z%(|eC+O;*t*3k#4QMqI7Zc;<^8W?6-@QCbSfGhrCeg%vb-5hptXu zg}LFk66A*-JlrfJA9JzEooE5Q1g+lc>af% zbfoy-fm1&k6~BU3R@BA&aJZzd*pFu-w?a)$?-P{Q*9d%@;W38^F*+4h_kyPjV_|H* zn%!A4f4jKnj(z4pcg+X>ba1bj*`Y}#8JV*NQNM&N&M#p-(Iy17IYp#tFV?9+A>4~< z-G`--+828DZ;s*`3)|iRi-wm%n6=U%-kqKP6E?)zG(EA3!=tVdk@liHx`hsMTk zU-2AaJ;|gWt~()pA4ai50vdbQdTDtw#an)s%{^dW2rt!hD@2{2RtOfy2^DnT!;a+b zck+Z}dA1db3u}f@4$0Ke2nN{!PBFkmf59ww8Xggk(1j#Um0zf@+q;wvOUh7o)&YC; zAg>V+w&b#%+`^CkiVyRmt(Hb2$#2L9p`2Sx#|Ni6`z{sMJdctsV}rd9hw3|SNJT>; z_qmy}%nwy6hi5=ScoAz0QcOIA8KgL*9F$UectoN=`=QK+x!k&x+lQfpq!t8S|N0 zMd|~SWIsjFCX#GlM!nMW8c;Lje|M8N-7?|uH9O;GlPrSqgCL8a!nPmAkzHSz4rMcJ zO5%aRYfVN=NetdcT87H$GszJLW0AeRktqv6tqct=KBwjf-*ks|DqK5ENT=~?_p-c> z$44xTx4n9=_RDk?pHOD?Je3uF%Kn1!=JCX}VGfy4SEf26`h<|8!OamKe=IHCgW8QD z+)0_+VPCQ;(1cRg)?49QTZ)PJy`yZFPfkH8A6i|9nr3gtci^w@C*#pO!L0a1}ztE``{mGU-btWkzkA-hFDLO zgLY>zpNIwIUq)ab1l|khe?K9gEKf9$BdN$QY#SXs$A6fM1^1QFNX5~W4#$7|Eb^Vc z=H^N#_{Z6~Xdi7Kqhp4hg2rS5-lv&M#*(q*R|yZtoJg6nxAJhr$dl!RG#%Pag5L#h z2~1SXRS%qO!K?)7#6 zceB4hoK{U50X4p>qhV(9St-Sr;*+n2FG;3U+zLB$uwu~a20JEN+9rlAzQz4WuOY0_ z(S)I4l5kR_e|ik+g*tvweZmM@VyK)w;^TbW2o*6DQ>yAThIkbO%?gq%D`q!tuegt1 zOnjt&SntOze970>?5Nqk>BHVbt5ja$6A~ew8DPj_L06e8;#F-j#YD38k(H14%$BZx z4tmNkm`=rgtZyp`Nl;81x5&;(=i1?1VGgr=-JZEppisRsF6Y zv+T!6g_~DUQwbD??@fUuVyGIBpcg5IMM4r%s3DXD7EuvVupmJVh9aITABlFr0MHh7e^cs z-fPH_O6<2Dg)w(}LGCvVce{<_J3D^0<7)fY9q#^}CtjC4pI-B{OLEX)(*$98Ww_;Z zyLM?|C*KThx)5F*C04eoF;Cr|)vE4)6p#A7ziJu2eR$1C{W5M$|B>+2UjDYr;+AX| zhtLOPv%+P9e^%X5C7(m|#WQ36KcR~*KckCI9G8-kP#ci;oi3*7hHUJ*$0`{Kb%YC- zzpS6@MpnG5m(h>aKS;>jd);O3tyMQ!P}QVbNAauHAru(u6aXUd%0g8K1OYe`5{P<~ zUsw0M&lEG0;6}fTZ8u)*XDLF^@(+0IE(M!K(^=H+X9>FQzybbb;Mj5OudS$q038Y% zN`O=wm!8*|&DCN;KHy5RhE3K!yDpuc{FLK%U#Qrpd4QSpng<~>Gdp>2KOjo2UPVR; zP&19Dq-wW6S7M5Q@=7R?(lW8aZhpL0NLzlhvXaZ?t`w%;#;IO!IT(TKLqDzs0?&)9 zFLB!Old-$Az)p*WJxs%$@mVJqk=11nO|TE=vpnR$H4x0>$zX9@KB*1=0pGA$w>)MP z)blw1MA@((26F(ItqT}|YWtpZ!x062PkI}LILzQUC|GEey#P$fAl5WyxImar_XcQf zBP4?&-UY>!Fo8w}CC#YJyF?u^#T?zxiASF0<)$@x(8PvJ@>HzgRYbI+>soP2Fu(4; zkIWQiPW_>2@2d#HNqW`32>FJ&(O|aT`~hsX)mfpkC2@M%faTXTalq@&8;ww5I{J$!deorg|`F-`jOd#GrW(z+m^S6Ps$PXTj;N*=NYzK0OMAT{4YOOX@9 zNRKR+gqrGlIY;V4rm}nbaY|3z>g0M_*plH3imf~ir`>3wEakI&U83V0G8?n2A&f<_ zI$Kk;RHtVt_*V0j=oj0NAzvp*)N3Ae?D!_)AQw3pGM?ov?^lk~Iw{bkGCfS@)vZW1 zWR{*%h6i2Sn5)?mYJvB1Q5&5u&#?x^M)0#%Efp+@4Tu9GH;rIEtJg+lRpX2fH|7Eg zCmm2oSw9e=y9gD4^0jEJP+n_Bp6E?Y&oryXI9P=U*@1L0Z|&XYOv1z&@Nbkw5<C*GD3eU-rL+w%+Kh| zlX+}8ggyGzHsVD_D|%jZ^eNw2rS&1$GaFVI%Sjkj_#`N$n=yB#HBSJCUtPB$Ho z+F3P_SkxPE#&d(yZ*}}P0~%?-K!)^G;e=C6q$D_l|KS%6y@?R_<2SDOVZBIBiA+jK UOA