def backtracking(filename):
###
# use backtracking to solve sudoku puzzle here,
# return the solution in the form of list of
# list as describe in the PDF with # of consistency
# checks done
###
game = utils.game(filename)
def recursiveBacktracking(game):
# consistencyChecks is incremented every time before checking for completion
game.consistencyChecks += 1
if game.isAssignmentComplete():
return True
pos = game.getUnassignedPositions(first=True) # Only get the first empty cell
x = pos[0][0]
y = pos[0][1]
validMoves = game.getValidMoves(x,y)
for move in validMoves:
game.sudokuBoard[x][y] = move
result = recursiveBacktracking(game)
if result is not None:
return result
game.sudokuBoard[x][y] = 0 # Failed assignment. Reset and return None.
return
if recursiveBacktracking(game):
return (game.sudokuBoard,game.consistencyChecks)
else:
return ("Solution not reached",game.consistencyChecks)
def minConflict(filename):
###
# use minConflict to solve sudoku puzzle here,
# return the solution in the form of list of
# list as describe in the PDF with # of consistency
# checks done
###
game = utils.game(filename)
maxSteps = 50000 # The maximum number of iterations
# Find all empty cells. These are the CSP's variables.
variables = [(i, j) for i in range(len(game.sudokuBoard))
for j in range(len(game.sudokuBoard))
if game.sudokuBoard[i][j] == 0]
# Initial assignment
for var in variables:
game.sudokuBoard[var[0]][var[1]] = game.leastConflicting(*var)
for i in range(maxSteps):
game.consistencyChecks += 1
if game.isAssignmentValid():
return (game.sudokuBoard, game.consistencyChecks)
# Choose one variable at random, and assign the least conflicting value
var = random.choice(variables)
game.sudokuBoard[var[0]][var[1]] = game.leastConflicting(*var)
# Out of iterations. Could not reach a valid assignment.
return ("Solution not reached", game.consistencyChecks)
def backtrackingMRVfwd(filename):
###
# use backtracking +MRV + forward propogation
# to solve sudoku puzzle here,
# return the solution in the form of list of
# list as describe in the PDF with # of consistency
# checks done
###
game = utils.game(filename)
def recursivebacktrackingMRVfwd(game):
game.consistencyChecks += 1
if game.isAssignmentComplete():
return True
pos = game.getUnassignedPositions()
mrvPos = game.getMRVPos(pos) # Find the most constrained position
x = pos[mrvPos][0]
y = pos[mrvPos][1]
moves = game.getValidMoves(x,y)
# Determine the least constraining value
constraintCount = list()
for i in range(len(moves)):
game.sudokuBoard[x][y] = moves[i]
constraintCount.append((moves[i],game.getConstraintCount(x,y)))
lcvMoves = sorted(constraintCount, key=lambda constraintCount:constraintCount[1],reverse = True)
# Backtrack
for move in lcvMoves:
game.sudokuBoard[x][y] = move[0]
if game.forwardChecking(x,y): # Perform forward checking
result = recursivebacktrackingMRVfwd(game)
if result is not None:
return result
game.sudokuBoard[x][y] = 0
return
if recursivebacktrackingMRVfwd(game):
return (game.sudokuBoard,game.consistencyChecks)
else:
return ("Solution not reached",game.consistencyChecks)
