package aima.gui.swing.applications.search.games;
import java.awt.Color;
import java.awt.Font;
import java.awt.GridLayout;
import java.awt.Insets;
import java.awt.event.ActionEvent;
import java.awt.event.ActionListener;
import java.util.ArrayList;
import java.util.Iterator;
import java.util.List;
import java.util.Properties;
import javax.swing.JButton;
import aima.core.agent.Action;
import aima.core.agent.Agent;
import aima.core.agent.Environment;
import aima.core.agent.Percept;
import aima.core.agent.impl.AbstractEnvironment;
import aima.core.environment.nqueens.AttackingPairsHeuristic;
import aima.core.environment.nqueens.NQueensBoard;
import aima.core.environment.nqueens.NQueensFunctionFactory;
import aima.core.environment.nqueens.NQueensGoalTest;
import aima.core.environment.nqueens.QueenAction;
import aima.core.search.framework.SearchAgent;
import aima.core.search.framework.SearchForActions;
import aima.core.search.framework.problem.ActionsFunction;
import aima.core.search.framework.problem.Problem;
import aima.core.search.framework.qsearch.GraphSearch;
import aima.core.search.framework.qsearch.TreeSearch;
import aima.core.search.informed.AStarSearch;
import aima.core.search.local.HillClimbingSearch;
import aima.core.search.local.Scheduler;
import aima.core.search.local.SimulatedAnnealingSearch;
import aima.core.search.uninformed.BreadthFirstSearch;
import aima.core.search.uninformed.DepthFirstSearch;
import aima.core.search.uninformed.DepthLimitedSearch;
import aima.core.search.uninformed.IterativeDeepeningSearch;
import aima.core.util.datastructure.XYLocation;
import aima.gui.swing.framework.AgentAppController;
import aima.gui.swing.framework.AgentAppEnvironmentView;
import aima.gui.swing.framework.AgentAppFrame;
import aima.gui.swing.framework.MessageLogger;
import aima.gui.swing.framework.SimpleAgentApp;
import aima.gui.swing.framework.SimulationThread;
/**
* Graphical n-queens game application. It demonstrates the performance of
* different search algorithms. An incremental problem formulation is supported
* as well as a complete-state formulation. Additionally, the user can make
* experiences with manual search.
*
* @author Ruediger Lunde
*/
public class NQueensApp extends SimpleAgentApp {
/** List of supported search algorithm names. */
protected static List<String> SEARCH_NAMES = new ArrayList<String>();
/** List of supported search algorithms. */
protected static List<SearchForActions> SEARCH_ALGOS = new ArrayList<SearchForActions>();
/** Adds a new item to the list of supported search algorithms. */
public static void addSearchAlgorithm(String name, SearchForActions algo) {
SEARCH_NAMES.add(name);
SEARCH_ALGOS.add(algo);
}
static {
addSearchAlgorithm("Depth First Search (Graph Search)",
new DepthFirstSearch(new GraphSearch()));
addSearchAlgorithm("Breadth First Search (Tree Search)",
new BreadthFirstSearch(new TreeSearch()));
addSearchAlgorithm("Breadth First Search (Graph Search)",
new BreadthFirstSearch(new GraphSearch()));
addSearchAlgorithm("Depth Limited Search (8)", new DepthLimitedSearch(8));
addSearchAlgorithm("Iterative Deepening Search", new IterativeDeepeningSearch());
addSearchAlgorithm("A* search (attacking pair heuristic)",
new AStarSearch(new GraphSearch(), new AttackingPairsHeuristic()));
addSearchAlgorithm("Hill Climbing Search", new HillClimbingSearch(new AttackingPairsHeuristic()));
addSearchAlgorithm("Simulated Annealing Search",
new SimulatedAnnealingSearch(new AttackingPairsHeuristic(), new Scheduler(20, 0.045, 1000)));
}
/** Returns a <code>NQueensView</code> instance. */
public AgentAppEnvironmentView createEnvironmentView() {
return new NQueensView();
}
/** Returns a <code>NQueensFrame</code> instance. */
@Override
public AgentAppFrame createFrame() {
return new NQueensFrame();
}
/** Returns a <code>NQueensController</code> instance. */
@Override
public AgentAppController createController() {
return new NQueensController();
}
// ///////////////////////////////////////////////////////////////
// main method
/**
* Starts the application.
*/
public static void main(String args[]) {
new NQueensApp().startApplication();
}
// ///////////////////////////////////////////////////////////////
// some inner classes
/**
* Adds some selectors to the base class and adjusts its size.
*/
protected static class NQueensFrame extends AgentAppFrame {
private static final long serialVersionUID = 1L;
public static String ENV_SEL = "EnvSelection";
public static String PROBLEM_SEL = "ProblemSelection";
public static String SEARCH_SEL = "SearchSelection";
public NQueensFrame() {
setTitle("N-Queens Application");
setSelectors(new String[] { ENV_SEL, PROBLEM_SEL, SEARCH_SEL },
new String[] { "Select Environment", "Select Problem Formulation", "Select Search" });
setSelectorItems(ENV_SEL, new String[] { "4 Queens", "8 Queens", "16 Queens", "32 Queens" }, 1);
setSelectorItems(PROBLEM_SEL, new String[] { "Incremental", "Complete-State" }, 0);
setSelectorItems(SEARCH_SEL, (String[]) SEARCH_NAMES.toArray(new String[] {}), 0);
setEnvView(new NQueensView());
setSize(800, 600);
}
}
/**
* Displays the informations provided by a <code>NQueensEnvironment</code>
* on a panel.
*/
protected static class NQueensView extends AgentAppEnvironmentView implements ActionListener {
private static final long serialVersionUID = 1L;
protected JButton[] squareButtons;
protected int currSize = -1;
protected NQueensView() {
}
@Override
public void setEnvironment(Environment env) {
super.setEnvironment(env);
showState();
}
@Override
public void agentAdded(Agent agent, Environment source) {
showState();
}
/** Agent value null indicates a user initiated action. */
@Override
public void agentActed(Agent agent, Percept percept, Action action, Environment source) {
showState();
notify((agent == null ? "User: " : "") + action.toString());
}
/**
* Displays the board state by labeling and coloring the square buttons.
*/
protected void showState() {
NQueensBoard board = ((NQueensEnvironment) env).getBoard();
if (currSize != board.getSize()) {
currSize = board.getSize();
removeAll();
setLayout(new GridLayout(currSize, currSize));
squareButtons = new JButton[currSize * currSize];
for (int i = 0; i < currSize * currSize; i++) {
JButton square = new JButton("");
square.setMargin(new Insets(0, 0, 0, 0));
square.setBackground((i % currSize) % 2 == (i / currSize) % 2 ? Color.WHITE : Color.LIGHT_GRAY);
square.addActionListener(this);
squareButtons[i] = square;
add(square);
}
}
for (int i = 0; i < currSize * currSize; i++)
squareButtons[i].setText("");
Font f = new java.awt.Font(Font.SANS_SERIF, Font.PLAIN,
Math.min(getWidth(), getHeight()) * 3 / 4 / currSize);
for (XYLocation loc : board.getQueenPositions()) {
JButton square = squareButtons[loc.getXCoOrdinate() + loc.getYCoOrdinate() * currSize];
square.setForeground(board.isSquareUnderAttack(loc) ? Color.RED : Color.BLACK);
square.setFont(f);
square.setText("Q");
}
validate();
}
/**
* When the user presses square buttons the board state is modified
* accordingly.
*/
@Override
public void actionPerformed(ActionEvent ae) {
for (int i = 0; i < currSize * currSize; i++) {
if (ae.getSource() == squareButtons[i]) {
NQueensController contr = (NQueensController) getController();
XYLocation loc = new XYLocation(i % currSize, i / currSize);
contr.modifySquare(loc);
}
}
}
}
/**
* Defines how to react on standard simulation button events.
*/
protected static class NQueensController extends AgentAppController {
protected NQueensEnvironment env = null;
protected SearchAgent agent = null;
protected boolean boardDirty;
/** Prepares next simulation. */
@Override
public void clear() {
prepare(null);
}
/**
* Creates an n-queens environment and clears the current search agent.
*/
@Override
public void prepare(String changedSelector) {
AgentAppFrame.SelectionState selState = frame.getSelection();
NQueensBoard board = null;
switch (selState.getIndex(NQueensFrame.ENV_SEL)) {
case 0: // 4 x 4 board
board = new NQueensBoard(4);
break;
case 1: // 8 x 8 board
board = new NQueensBoard(8);
break;
case 2: // 8 x 8 board
board = new NQueensBoard(16);
break;
case 3: // 32 x 32 board
board = new NQueensBoard(32);
break;
}
env = new NQueensEnvironment(board);
if (selState.getIndex(NQueensFrame.PROBLEM_SEL) == 1)
for (int i = 0; i < board.getSize(); i++)
board.addQueenAt(new XYLocation(i, 0));
boardDirty = false;
agent = null;
frame.getEnvView().setEnvironment(env);
}
/**
* Creates a new search agent and adds it to the current environment if
* necessary.
*/
protected void addAgent() throws Exception {
if (agent != null && agent.isDone()) {
env.removeAgent(agent);
agent = null;
}
if (agent == null) {
int pSel = frame.getSelection().getIndex(NQueensFrame.PROBLEM_SEL);
int sSel = frame.getSelection().getIndex(NQueensFrame.SEARCH_SEL);
ActionsFunction af;
if (pSel == 0)
af = NQueensFunctionFactory.getIActionsFunction();
else
af = NQueensFunctionFactory.getCActionsFunction();
Problem problem = new Problem(env.getBoard(), af, NQueensFunctionFactory.getResultFunction(),
new NQueensGoalTest());
SearchForActions search = SEARCH_ALGOS.get(sSel);
agent = new SearchAgent(problem, search);
env.addAgent(agent);
}
}
/** Checks whether simulation can be started. */
@Override
public boolean isPrepared() {
int problemSel = frame.getSelection().getIndex(NQueensFrame.PROBLEM_SEL);
return problemSel == 1 || (agent == null || !agent.isDone())
&& (!boardDirty || env.getBoard().getNumberOfQueensOnBoard() == 0);
}
/** Starts simulation. */
@Override
public void run(MessageLogger logger) {
logger.log("<simulation-log>");
try {
addAgent();
while (!agent.isDone() && !frame.simulationPaused()) {
Thread.sleep(200);
env.step();
}
} catch (InterruptedException e) {
// nothing to do...
} catch (Exception e) {
e.printStackTrace(); // probably search has failed...
}
logger.log(getStatistics());
logger.log("</simulation-log>\n");
}
/** Executes one simulation step. */
@Override
public void step(MessageLogger logger) {
try {
addAgent();
env.step();
} catch (Exception e) {
e.printStackTrace(); // probably search has failed...
}
}
/** Updates the status of the frame after simulation has finished. */
public void update(SimulationThread simulationThread) {
if (simulationThread.isCanceled()) {
frame.setStatus("Task canceled.");
} else if (frame.simulationPaused()) {
frame.setStatus("Task paused.");
} else {
frame.setStatus("Task completed.");
}
}
/** Provides a text with statistical information about the last run. */
private String getStatistics() {
StringBuffer result = new StringBuffer();
Properties properties = agent.getInstrumentation();
Iterator<Object> keys = properties.keySet().iterator();
while (keys.hasNext()) {
String key = (String) keys.next();
String property = properties.getProperty(key);
result.append("\n" + key + " : " + property);
}
return result.toString();
}
public void modifySquare(XYLocation loc) {
boardDirty = true;
String atype;
if (env.getBoard().queenExistsAt(loc))
atype = QueenAction.REMOVE_QUEEN;
else
atype = QueenAction.PLACE_QUEEN;
env.executeAction(null, new QueenAction(atype, loc));
agent = null;
frame.updateEnabledState();
}
}
/** Simple environment maintaining just the current board state. */
public static class NQueensEnvironment extends AbstractEnvironment {
NQueensBoard board;
public NQueensEnvironment(NQueensBoard board) {
this.board = board;
}
public NQueensBoard getBoard() {
return board;
}
/**
* Executes the provided action and returns null.
*/
@Override
public void executeAction(Agent agent, Action action) {
if (action instanceof QueenAction) {
QueenAction act = (QueenAction) action;
XYLocation loc = new XYLocation(act.getX(), act.getY());
if (act.getName() == QueenAction.PLACE_QUEEN)
board.addQueenAt(loc);
else if (act.getName() == QueenAction.REMOVE_QUEEN)
board.removeQueenFrom(loc);
else if (act.getName() == QueenAction.MOVE_QUEEN)
board.moveQueenTo(loc);
}
}
/** Returns null. */
@Override
public Percept getPerceptSeenBy(Agent anAgent) {
return null;
}
}
}