All Possible Tic Tac Toe Winning Combinations - algorithm

I had an interview were I was asked a seemingly simple algorithm question: "Write an algorithm to return me all possible winning combinations for tic tac toe." I still can't figure out an efficient way to handle this. Is there a standard algorithm or common that should be applied to similar questions like this that I'm not aware of?

This is one of those problems that's actually simple enough for brute force and, while you could use combinatorics, graph theory, or many other complex tools to solve it, I'd actually be impressed by applicants that recognise the fact there's an easier way (at least for this problem).
There are only 39, or 19,683 possible combinations of placing x, o or <blank> in the grid, and not all of those are valid.
First, a valid game position is one where the difference between x and o counts is no more than one, since they have to alternate moves.
In addition, it's impossible to have a state where both sides have three in a row, so they can be discounted as well. If both have three in a row, then one of them would have won in the previous move.
There's actually another limitation in that it's impossible for one side to have won in two different ways without a common cell (again, they would have won in a previous move), meaning that:
XXX
OOO
XXX
cannot be achieved, while:
XXX
OOX
OOX
can be. But we can actually ignore that since there's no way to win two ways without a common cell without having already violated the "maximum difference of one" rule, since you need six cells for that, with the opponent only having three.
So I would simply use brute force and, for each position where the difference is zero or one between the counts, check the eight winning possibilities for both sides. Assuming only one of them has a win, that's a legal, winning game.
Below is a proof of concept in Python, but first the output of time when run on the process sending output to /dev/null to show how fast it is:
real 0m0.169s
user 0m0.109s
sys 0m0.030s
The code:
def won(c, n):
if c[0] == n and c[1] == n and c[2] == n: return 1
if c[3] == n and c[4] == n and c[5] == n: return 1
if c[6] == n and c[7] == n and c[8] == n: return 1
if c[0] == n and c[3] == n and c[6] == n: return 1
if c[1] == n and c[4] == n and c[7] == n: return 1
if c[2] == n and c[5] == n and c[8] == n: return 1
if c[0] == n and c[4] == n and c[8] == n: return 1
if c[2] == n and c[4] == n and c[6] == n: return 1
return 0
pc = [' ', 'x', 'o']
c = [0] * 9
for c[0] in range (3):
for c[1] in range (3):
for c[2] in range (3):
for c[3] in range (3):
for c[4] in range (3):
for c[5] in range (3):
for c[6] in range (3):
for c[7] in range (3):
for c[8] in range (3):
countx = sum([1 for x in c if x == 1])
county = sum([1 for x in c if x == 2])
if abs(countx-county) < 2:
if won(c,1) + won(c,2) == 1:
print " %s | %s | %s" % (pc[c[0]],pc[c[1]],pc[c[2]])
print "---+---+---"
print " %s | %s | %s" % (pc[c[3]],pc[c[4]],pc[c[5]])
print "---+---+---"
print " %s | %s | %s" % (pc[c[6]],pc[c[7]],pc[c[8]])
print
As one commenter has pointed out, there is one more restriction. The winner for a given board cannot have less cells than the loser since that means the loser just moved, despite the fact the winner had already won on the last move.
I won't change the code to take that into account but it would be a simple matter of checking who has the most cells (the last person that moved) and ensuring the winning line belonged to them.

Another way could be to start with each of the eight winning positions,
xxx ---
--- xxx
--- --- ... etc.,
and recursively fill in all legal combinations (start with inserting 2 o's, then add an x for each o ; avoid o winning positions):
xxx xxx xxx
oo- oox oox
--- o-- oox ... etc.,

Today I had an interview with Apple and I had the same question. I couldn't think well at that moment. Later one on, before going to a meeting I wrote the function for the combinations in 15 minutes, and when I came back from the meeting I wrote the validation function again in 15 minutes. I get nervous at interviews, Apple not trusts my resume, they only trust what they see in the interview, I don't blame them, many companies are the same, I just say that something in this hiring process doesn't look quite smart.
Anyways, here is my solution in Swift 4, there are 8 lines of code for the combinations function and 17 lines of code to check a valid board.
Cheers!!!
// Not used yet: 0
// Used with x : 1
// Used with 0 : 2
// 8 lines code to get the next combination
func increment ( _ list: inout [Int], _ base: Int ) -> Bool {
for digit in 0..<list.count {
list[digit] += 1
if list[digit] < base { return true }
list[digit] = 0
}
return false
}
let incrementTicTacToe = { increment(&$0, 3) }
let win0_ = [0,1,2] // [1,1,1,0,0,0,0,0,0]
let win1_ = [3,4,5] // [0,0,0,1,1,1,0,0,0]
let win2_ = [6,7,8] // [0,0,0,0,0,0,1,1,1]
let win_0 = [0,3,6] // [1,0,0,1,0,0,1,0,0]
let win_1 = [1,4,7] // [0,1,0,0,1,0,0,1,0]
let win_2 = [2,5,8] // [0,0,1,0,0,1,0,0,1]
let win00 = [0,4,8] // [1,0,0,0,1,0,0,0,1]
let win11 = [2,4,6] // [0,0,1,0,1,0,1,0,0]
let winList = [ win0_, win1_, win2_, win_0, win_1, win_2, win00, win11]
// 16 lines to check a valid board, wihtout countin lines of comment.
func winCombination (_ tictactoe: [Int]) -> Bool {
var count = 0
for win in winList {
if tictactoe[win[0]] == tictactoe[win[1]],
tictactoe[win[1]] == tictactoe[win[2]],
tictactoe[win[2]] != 0 {
// If the combination exist increment count by 1.
count += 1
}
if count == 2 {
return false
}
}
var indexes = Array(repeating:0, count:3)
for num in tictactoe { indexes[num] += 1 }
// '0' and 'X' must be used the same times or with a diference of one.
// Must one and only one valid combination
return abs(indexes[1] - indexes[2]) <= 1 && count == 1
}
// Test
var listToIncrement = Array(repeating:0, count:9)
var combinationsCount = 1
var winCount = 0
while incrementTicTacToe(&listToIncrement) {
if winCombination(listToIncrement) == true {
winCount += 1
}
combinationsCount += 1
}
print("There is \(combinationsCount) combinations including possible and impossible ones.")
print("There is \(winCount) combinations for wining positions.")
/*
There are 19683 combinations including possible and impossible ones.
There are 2032 combinations for winning positions.
*/
listToIncrement = Array(repeating:0, count:9)
var listOfIncremented = ""
for _ in 0..<1000 { // Win combinations for the first 1000 combinations
_ = incrementTicTacToe(&listToIncrement)
if winCombination(listToIncrement) == true {
listOfIncremented += ", \(listToIncrement)"
}
}
print("List of combinations: \(listOfIncremented)")
/*
List of combinations: , [2, 2, 2, 1, 1, 0, 0, 0, 0], [1, 1, 1, 2, 2, 0, 0, 0, 0],
[2, 2, 2, 1, 0, 1, 0, 0, 0], [2, 2, 2, 0, 1, 1, 0, 0, 0], [2, 2, 0, 1, 1, 1, 0, 0, 0],
[2, 0, 2, 1, 1, 1, 0, 0, 0], [0, 2, 2, 1, 1, 1, 0, 0, 0], [1, 1, 1, 2, 0, 2, 0, 0, 0],
[1, 1, 1, 0, 2, 2, 0, 0, 0], [1, 1, 0, 2, 2, 2, 0, 0, 0], [1, 0, 1, 2, 2, 2, 0, 0, 0],
[0, 1, 1, 2, 2, 2, 0, 0, 0], [1, 2, 2, 1, 0, 0, 1, 0, 0], [2, 2, 2, 1, 0, 0, 1, 0, 0],
[2, 2, 1, 0, 1, 0, 1, 0, 0], [2, 2, 2, 0, 1, 0, 1, 0, 0], [2, 2, 2, 1, 1, 0, 1, 0, 0],
[2, 0, 1, 2, 1, 0, 1, 0, 0], [0, 2, 1, 2, 1, 0, 1, 0, 0], [2, 2, 1, 2, 1, 0, 1, 0, 0],
[1, 2, 0, 1, 2, 0, 1, 0, 0], [1, 0, 2, 1, 2, 0, 1, 0, 0], [1, 2, 2, 1, 2, 0, 1, 0, 0],
[2, 2, 2, 0, 0, 1, 1, 0, 0]
*/

This is a java equivalent code sample
package testit;
public class TicTacToe {
public static void main(String[] args) {
// TODO Auto-generated method stub
// 0 1 2
// 3 4 5
// 6 7 8
char[] pc = {' ' ,'o', 'x' };
char[] c = new char[9];
// initialize c
for (int i = 0; i < 9; i++)
c[i] = pc[0];
for (int i = 0; i < 3; i++) {
c[0] = pc[i];
for (int j = 0; j < 3; j++) {
c[1] = pc[j];
for (int k = 0; k < 3; k++) {
c[2] = pc[k];
for (int l = 0; l < 3; l++) {
c[3] = pc[l];
for (int m = 0; m < 3; m++) {
c[4] = pc[m];
for (int n = 0; n < 3; n++) {
c[5] = pc[n];
for (int o = 0; o < 3; o++) {
c[6] = pc[o];
for (int p = 0; p < 3; p++) {
c[7] = pc[p];
for (int q = 0; q < 3; q++) {
c[8] = pc[q];
int countx = 0;
int county = 0;
for(int r = 0 ; r<9 ; r++){
if(c[r] == 'x'){
countx = countx + 1;
}
else if(c[r] == 'o'){
county = county + 1;
}
}
if(Math.abs(countx - county) < 2){
if(won(c, pc[2])+won(c, pc[1]) == 1 ){
System.out.println(c[0] + " " + c[1] + " " + c[2]);
System.out.println(c[3] + " " + c[4] + " " + c[5]);
System.out.println(c[6] + " " + c[7] + " " + c[8]);
System.out.println("*******************************************");
}
}
}
}
}
}
}
}
}
}
}
}
public static int won(char[] c, char n) {
if ((c[0] == n) && (c[1] == n) && (c[2] == n))
return 1;
else if ((c[3] == n) && (c[4] == n) && (c[5] == n))
return 1;
else if ((c[6] == n) && (c[7] == n) && (c[8] == n))
return 1;
else if ((c[0] == n) && (c[3] == n) && (c[6] == n))
return 1;
else if ((c[1] == n) && (c[4] == n) && (c[7] == n))
return 1;
else if ((c[2] == n) && (c[5] == n) && (c[8] == n))
return 1;
else if ((c[0] == n) && (c[4] == n) && (c[8] == n))
return 1;
else if ((c[2] == n) && (c[4] == n) && (c[6] == n))
return 1;
else
return 0;
}
}
`

Below Solution generates all possible combinations using recursion
It has eliminated impossible combinations and returned 888 Combinations
Below is a working code Possible winning combinations of the TIC TAC TOE game
const players = ['X', 'O'];
let gameBoard = Array.from({ length: 9 });
const winningCombination = [
[ 0, 1, 2 ],
[ 3, 4, 5 ],
[ 6, 7, 8 ],
[ 0, 3, 6 ],
[ 1, 4, 7 ],
[ 2, 5, 8 ],
[ 0, 4, 8 ],
[ 2, 4, 6 ],
];
const isWinningCombination = (board)=> {
if((Math.abs(board.filter(a => a === players[0]).length -
board.filter(a => a === players[1]).length)) > 1) {
return false
}
let winningComb = 0;
players.forEach( player => {
winningCombination.forEach( combinations => {
if (combinations.every(combination => board[combination] === player )) {
winningComb++;
}
});
});
return winningComb === 1;
}
const getCombinations = (board) => {
let currentBoard = [...board];
const firstEmptySquare = board.indexOf(undefined)
if (firstEmptySquare === -1) {
return isWinningCombination(board) ? [board] : [];
} else {
return [...players, ''].reduce((prev, next) => {
currentBoard[firstEmptySquare] = next;
if(next !== '' && board.filter(a => a === next).length > (gameBoard.length / players.length)) {
return [...prev]
}
return [board, ...prev, ...getCombinations(currentBoard)]
}, [])
}
}
const startApp = () => {
let combination = getCombinations(gameBoard).filter(board =>
board.every(item => !(item === undefined)) && isWinningCombination(board)
)
printCombination(combination)
}
const printCombination = (combination)=> {
const ulElement = document.querySelector('.combinations');
combination.forEach(comb => {
let node = document.createElement("li");
let nodePre = document.createElement("pre");
let textnode = document.createTextNode(JSON.stringify(comb));
nodePre.appendChild(textnode);
node.appendChild(nodePre);
ulElement.appendChild(node);
})
}
startApp();

This discovers all possible combinations for tic tac toe (255,168) -- written in JavaScript using recursion. It is not optimized, but gets you what you need.
const [EMPTY, O, X] = [0, 4, 1]
let count = 0
let coordinate = [
EMPTY, EMPTY, EMPTY,
EMPTY, EMPTY, EMPTY,
EMPTY, EMPTY, EMPTY
]
function reducer(arr, sumOne, sumTwo = null) {
let func = arr.reduce((sum, a) => sum + a, 0)
if((func === sumOne) || (func === sumTwo)) return true
}
function checkResult() {
let [a1, a2, a3, b1, b2, b3, c1, c2, c3] = coordinate
if(reducer([a1,a2,a3], 3, 12)) return true
if(reducer([a1,b2,c3], 3, 12)) return true
if(reducer([b1,b2,b3], 3, 12)) return true
if(reducer([c1,c2,c3], 3, 12)) return true
if(reducer([a3,b2,c1], 3, 12)) return true
if(reducer([a1,b1,c1], 3, 12)) return true
if(reducer([a2,b2,c2], 3, 12)) return true
if(reducer([a3,b3,c3], 3, 12)) return true
if(reducer([a1,a2,a3,b1,b2,b3,c1,c2,c3], 21)) return true
return false
}
function nextPiece() {
let [countX, countO] = [0, 0]
for(let i = 0; i < coordinate.length; i++) {
if(coordinate[i] === X) countX++
if(coordinate[i] === O) countO++
}
return countX === countO ? X : O
}
function countGames() {
if (checkResult()) {
count++
}else {
for (let i = 0; i < 9; i++) {
if (coordinate[i] === EMPTY) {
coordinate[i] = nextPiece()
countGames()
coordinate[i] = EMPTY
}
}
}
}
countGames()
console.log(count)
I separated out the checkResult returns in case you want to output various win conditions.

Could be solved with brute force but keep in mind the corner cases like player2 can't move when player1 has won and vice versa. Also remember Difference between moves of player1 and player can't be greater than 1 and less than 0.
I have written code for validating whether provided combination is valid or not, might soon post on github.

Related

How to solve flow game using Google OR tools?

I tried to make solver for flow game using google-OR tools.
I made a few rules for the corner to only contains corner pipes, but other than that, i can not figure out how to make the pipe connected to each other nor how to tell the model to make a pipe that is connecting to each other.
A few snippet
pipe_types = {
0: " ",
1: "-",
2: "|",
3: "┗" ,
4: "┛" ,
5: "┓",
6: "┏",
7: "●"
}
model = cp_model.CpModel()
filled_map = [[0,0,0,0],
[0,0,7,0],
[0,0,0,0],
[0,7,0,0]]
mesh_size = int(np.sqrt(len(np.array(filled_map).flatten())))
target_map = [[model.NewIntVar(1, 6, 'column: %i' % i) for i in range(mesh_size)] for j in range(mesh_size)]
flow_map = init_map(model, target_map, filled_map)
for i in range(len(flow_map)):
for j in range(len(flow_map[0])):
# check if top or bottom side
if (i == 0) or (i == len(flow_map)-1):
model.Add(flow_map[i][j] != 2)
# check if left or right side
if (j == 0) or (j == len(flow_map[0])-1):
model.Add(flow_map[i][j] != 1)
# left up corner
if (i == 0) & (j == 0):
model.Add(flow_map[i][j] != 3)
model.Add(flow_map[i][j] != 4)
model.Add(flow_map[i][j] != 5)
# right up corner
if (i == 0) & (j == len(flow_map[0])-1):
model.Add(flow_map[i][j] != 3)
model.Add(flow_map[i][j] != 4)
model.Add(flow_map[i][j] != 6)
# left bottom corner
if (i == len(flow_map)-1) & (j == 0):
model.Add(flow_map[i][j] != 4)
model.Add(flow_map[i][j] != 5)
model.Add(flow_map[i][j] != 6)
# right bottom corner
if (i == len(flow_map)-1) & (j == len(flow_map[0])-1):
model.Add(flow_map[i][j] != 3)
model.Add(flow_map[i][j] != 5)
model.Add(flow_map[i][j] != 6)
# Solving
status = solver.Solve(model)
res = []
if status == cp_model.OPTIMAL or status == cp_model.FEASIBLE:
for i in range(len(flow_map)):
for j in range(len(flow_map[0])):
res.append(solver.Value(flow_map[i][j]))
print(solver.Value(flow_map[i][j]), end=" ")
print()
This would results horizontal pipes on the center of the mesh. Later on, i would have to figure out how to add color and such on this too.
Is there any pointer on how to make this on OR tools?
Edit 1:
Based on David Eisenstat's answer, I can find solution. Visualizing this solution based on JohanC's answer, I get this result.
Can I get the pathing made from google-OR tools?
Edit 2:
Using hamilton path from "Hamiltonian" path using Python
I could generate somewhat correct pathing.
But it feels so weird since OR tools already calculate the pathing, and I have to recalculate the path. The path generated from "Hamiltonian" path using Python doesn't show all possible combinations. If I can take the path from OR tools, I think that would be my best interest.
As I don't have experience with OR-tools, here is an approach with Z3.
The initial board is represented by numbers for the end points, one number for each color. The idea is a bit similar to how Sudoku is represented.
Each other cell on the board will get either a value for zero, or a number. This number should be equal to exactly two of its neighbors.
The initial endpoints should have exactly one neighbor with its color.
from z3 import Solver, Sum, Int, If, And, Or, sat
def plot_solution(S):
import matplotlib.pyplot as plt
ax = plt.gca()
colors = plt.cm.tab10.colors
for i in range(M):
for j in range(N):
if board[i][j] != 0:
ax.scatter(j, i, s=500, color=colors[board[i][j]])
if S[i][j] != 0:
for k in range(M):
for l in range(N):
if abs(k - i) + abs(l - j) == 1 and S[i][j] == S[k][l]:
ax.plot([j, l], [i, k], color=colors[S[i][j]], lw=15)
ax.set_ylim(M - 0.5, -0.5)
ax.set_xlim(-0.5, N - 0.5)
ax.set_aspect('equal')
ax.set_facecolor('black')
ax.set_yticks([i + 0.5 for i in range(M - 1)], minor=True)
ax.set_xticks([j + 0.5 for j in range(N - 1)], minor=True)
ax.grid(b=True, which='minor', color='white')
ax.set_xticks([])
ax.set_yticks([])
ax.tick_params(axis='both', which='both', length=0)
plt.show()
board = [[1, 0, 0, 2, 3],
[0, 0, 0, 4, 0],
[0, 0, 4, 0, 0],
[0, 2, 3, 0, 5],
[0, 1, 5, 0, 0]]
M = len(board)
N = len(board[0])
B = [[Int(f'B_{i}_{j}') for j in range(N)] for i in range(M)]
s = Solver()
s.add(([If(board[i][j] != 0, B[i][j] == board[i][j], And(B[i][j] >= 0, B[i][j] < 10))
for j in range(N) for i in range(M)]))
for i in range(M):
for j in range(N):
same_neighs_ij = Sum([If(B[i][j] == B[k][l], 1, 0)
for k in range(M) for l in range(N) if abs(k - i) + abs(l - j) == 1])
if board[i][j] != 0:
s.add(same_neighs_ij == 1)
else:
s.add(Or(same_neighs_ij == 2, B[i][j] == 0))
if s.check() == sat:
m = s.model()
S = [[m[B[i][j]].as_long() for j in range(N)] for i in range(M)]
print(S)
plot_solution(S)
Solution:
[[1, 2, 2, 2, 3],
[1, 2, 4, 4, 3],
[1, 2, 4, 3, 3],
[1, 2, 3, 3, 5],
[1, 1, 5, 5, 5]]
As mentioned in the comments, a possible requirement is that all cells would need to be colored. This would need a more complicated approach. Here is an example of such a configuration for which the above code could create a solution that connects all end points without touching all cells:
board = [[0, 1, 2, 0, 0, 0, 0],
[1, 3, 4, 0, 3, 5, 0],
[0, 0, 0, 0, 0, 0, 0],
[0, 2, 0, 4, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0],
[0, 5, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0]]
The best way is probably with AddCircuit. This constraint takes a directed graph where each arc is labeled with a literal and requires that the arcs labeled true form a subgraph where each node has in- and out-degree 1, and further that there is at most one cycle that is not a self-loop. By forcing an arc from the end to the beginning, we can use this constraint type to require that there is a single path from the beginning to the end.
The documentation is somewhat poor, so here's a working code sample. I'll leave the drawing part to you.
import collections
from ortools.sat.python import cp_model
def validate_board_and_count_colors(board):
assert isinstance(board, list)
assert all(isinstance(row, list) for row in board)
assert len(set(map(len, board))) == 1
colors = collections.Counter(square for row in board for square in row)
del colors[0]
assert all(count == 2 for count in colors.values())
num_colors = len(colors)
assert set(colors.keys()) == set(range(1, num_colors + 1))
return num_colors
def main(board):
num_colors = validate_board_and_count_colors(board)
model = cp_model.CpModel()
solution = [
[square or model.NewIntVar(1, num_colors, "") for (j, square) in enumerate(row)]
for (i, row) in enumerate(board)
]
true = model.NewBoolVar("")
model.AddBoolOr([true])
for color in range(1, num_colors + 1):
endpoints = []
arcs = []
for i, row in enumerate(board):
for j, square in enumerate(row):
if square == color:
endpoints.append((i, j))
else:
arcs.append(((i, j), (i, j)))
if i < len(board) - 1:
arcs.append(((i, j), (i + 1, j)))
if j < len(row) - 1:
arcs.append(((i, j), (i, j + 1)))
(i1, j1), (i2, j2) = endpoints
k1 = i1 * len(row) + j1
k2 = i2 * len(row) + j2
arc_variables = [(k2, k1, true)]
for (i1, j1), (i2, j2) in arcs:
k1 = i1 * len(row) + j1
k2 = i2 * len(row) + j2
edge = model.NewBoolVar("")
if k1 == k2:
model.Add(solution[i1][j1] != color).OnlyEnforceIf(edge)
arc_variables.append((k1, k1, edge))
else:
model.Add(solution[i1][j1] == color).OnlyEnforceIf(edge)
model.Add(solution[i2][j2] == color).OnlyEnforceIf(edge)
forward = model.NewBoolVar("")
backward = model.NewBoolVar("")
model.AddBoolOr([edge, forward.Not()])
model.AddBoolOr([edge, backward.Not()])
model.AddBoolOr([edge.Not(), forward, backward])
model.AddBoolOr([forward.Not(), backward.Not()])
arc_variables.append((k1, k2, forward))
arc_variables.append((k2, k1, backward))
model.AddCircuit(arc_variables)
solver = cp_model.CpSolver()
status = solver.Solve(model)
if status == cp_model.OPTIMAL:
for row in solution:
print("".join(str(solver.Value(x)) for x in row))
if __name__ == "__main__":
main(
[
[1, 0, 0, 2, 3],
[0, 0, 0, 4, 0],
[0, 0, 4, 0, 0],
[0, 2, 3, 0, 5],
[0, 1, 5, 0, 0],
]
)

Decompose string to form a valid expression

I am given a string S (of integers) and a number N. I want to insert arbitrary number of '+' in S so that the sum becomes equal to N.
Ex:<br>
S = 15112 and N = 28<br>
Ans is : 15+11+2<br>
S = 120012 and N = 33<br>
Ans is : 1+20+012<br>
S = 123 and N = 123<br>
Ans is : 123
given : |S| <= 120 and N <= 10^6
It is guarenteed that S and N are given such that it is always possible to form valid expression. Is there any algorithm which can solve this? I tried to think on it but couldn't come up with solution.
There may be more efficient ways to do this, but since you have nothing so far…
You can simply find all combinations of a boolean array that indicates whether a plus should exist between the numbers or not.
For example: with an input of 112134, 1 + 12 + 13 + 4 can be represented with the boolean array [true, false, true, false, true] indicating that there is a plus after the 1st, 3rd, and 5th numbers. The problem then reduces to finding which combinations add to your number. There are lot of ways to find combinations. Recursive backtracking is a classic.
In javascript/node this might look like this:
function splitOnIndexes(arr, a) {
// split the array into numbers based on the booleans
let current = "" + arr[0]
let output = []
for (let i = 0; i < a.length; i++) {
if (!a[i]) {
current += arr[i + 1]
} else {
output.push(current)
current = "" + arr[i + 1]
}
}
output.push(current)
return output
}
function findSum(input, total) {
function backtrack(n, k = 0, a = []) {
const sum = (arr) => arr.reduce((a, c) => a + parseInt(c), 0)
if (k === n) {
let ans = splitOnIndexes(input, a)
if (sum(ans) === total) {
console.log(ans.join(' + '))
}
} else {
k = k + 1
let c = [true, false]
for (let i = 0; i < 2; i++) {
a[k - 1] = c[i]
backtrack(n, k, a)
}
}
}
backtrack(input.length - 1)
}
findSum('15112', 28)
findSum('120012', 33)
findSum('123', 123)
As you can see, more than one answer is possible. Your first example is solved with both 15+1+12 and 15+11+2. If you only need one, you can of course stop early.
The idea is to use dynamic programming, you only care about sums between 0 and 10^6 and only have 120 possible indexes. if dp[i][j] = x, it means that from index x of the string, we went to index i (so we added a + before i) and we got a sum of j. This leads to a O(|S| * N) solution:
#include <iostream>
#include <string>
#include <vector>
using namespace std;
string s;
long n;
long dp[123][1000001];
void solve (int index, long sum) {//index = what index of s still remains to scan. sum = the sum we have accumulated till now
if (sum >= n or index >= s.length()) return;
if (dp[index][sum] != -1) return;
if (index == n and sum == n) return;
long num = 0;
for (int i = 0; i < 7 && index + i < s.length(); i++) { //N has 6 digits at most
num = stoi(s.substr(index, i + 1));
solve(index + i + 1, sum + num);
if (sum + num <= n) {
dp[index + i + 1][sum + num] = index;
}
}
}
int main () {
cin >> s;
cin >> n;
for (int i = 0; i < 121; i++) {
for (int j = 0; j < 1000001; j++) {
dp[i][j] = -1;
}
}
solve(0, 0);
int sum = n;
int idx = s.length();
vector<string> nums;
//reconstruct solution
while (idx != 0) {
nums.push_back(s.substr(dp[idx][sum], idx - dp[idx][sum]));
idx = dp[idx][sum];
sum -= stoi(nums[nums.size() - 1]);
}
for (int i = nums.size() -1; i >= 0; i--) {
cout << nums[i];
if (i != 0) cout << "+";
}
}
This is a Ruby version with step by step explanation of the algorithm, so you can easily code in C++ (or I'll try later).
# Let's consider that we extracted the values from text, so we already have the string of int and the result as integer:
string_of_int = "15112"
result = 28
# The basic idea is to find a map (array) that tells how to group digits, for example
sum_map = [2, 1, 2]
# This means that string_of_int is mapped into the following numbers
# 15, 1, 12
# then sum the numbers, in this case 15+1+12 = 28
# For finding a the solution we need to map
# all the possible combinations of addition given the n digits of the string_of_int then check if the sum is equal to the result
# We call k the number of digits of string_of_int
# in ruby we can build an array called sum_maps
# containing all the possible permutations like this:
k = string_of_int.length # => 5
sum_maps = []
k.times do |length|
(1..k).to_a.repeated_permutation(length).each {|e| sum_maps << e if e.inject(:+) == k}
end
sum_maps
# => [[1, 5], [2, 4], [3, 3], [4, 2], [5, 1], [1, 1, 4], [1, 2, 3], [1, 3, 2], [1, 4, 1], [2, 1, 3], [2, 2, 2], [2, 3, 1], [3, 1, 2], [3, 2, 1], [4, 1, 1]]
# Now must check which of of the sum_map is giving us the required result.
#
# First, to keep the code short and DRY,
# better to define a couple of useful methods for the String class to use then:
class String
def group_digits_by(sum_map)
string_of_int_splitted = self.split("")
grouped_digits = []
sum_map.each { |n| grouped_digits << string_of_int_splitted.shift(n).join.to_i}
grouped_digits.reject { |element| element == 0 }
end
def sum_grouped_of_digits_by(sum_map)
group_digits_by(sum_map).inject(:+)
end
end
# So we can call the methods directly on the string
# for example, in ruby:
string_of_int.group_digits_by sum_map #=> [15, 1, 12]
string_of_int.sum_grouped_of_digits_by sum_map #=> 28
# Now that we have this metods, we just iterate through the sum_maps array
# and apply it for printing out the sm_map if the sum of grouped digits is equal to the result
# coded in ruby it is:
combinations = []
sum_maps.each { |sum_map| combinations << string_of_int.group_digits_by(sum_map) if string_of_int.sum_grouped_of_digits_by(sum_map) == result }
p combinations.uniq
# => [[15, 1, 12], [15, 11, 2]]
In short, written as a Ruby module it becomes:
module GuessAddition
class ::String
def group_digits_by(sum_map)
string_of_int_splitted = self.split("")
grouped_digits = []
sum_map.each { |n| grouped_digits << string_of_int_splitted.shift(n).join.to_i}
grouped_digits.reject { |element| element == 0 }
end
def sum_grouped_of_digits_by(sum_map)
group_digits_by(sum_map).inject(:+)
end
end
def self.guess_this(string_of_int, result)
k = string_of_int.length
sum_maps = []
k.times { |length| (1..k).to_a.repeated_permutation(length).each {|e| sum_maps << e if e.inject(:+) == k} }
combinations = []
sum_maps.each { |sum_map| combinations << string_of_int.group_digits_by(sum_map) if string_of_int.sum_grouped_of_digits_by(sum_map) == result }
combinations.uniq
end
end
p GuessAddition::guess_this("15112", 28) # => [[15, 1, 12], [15, 11, 2]]

What is the Grooviest way to implement an incrementing integer list?

I need a data structure which is a list of integers but each time an integer is added to it the value it stores is the sum of the values it contains plus the value that is added.
For example:
def incrementingList = []
incrementingList.add(5) // now it has 5
incrementingList.add(3) // now it has 5,8
incrementingList.add(2) // now it has 5,8,10
Is there a groovy way to implement this so it can be ready to use as in the example?
UPDATE
What if it is possible for this list to contain 0s and if the last element is a 0 then it should increment by the last non 0 element?
You can use metaprogramming to define custom method:
List.metaClass.plusAdd = { e ->
delegate.isEmpty() ? delegate.add(e) : delegate.add(delegate.last() + e)
}
def l = []
l.plusAdd(5)
l.plusAdd(3)
l.plusAdd(2)
assert l == [5, 8, 10]
EDIT
Update for adding last non-zero element:
List.metaClass.plusAdd = { e ->
if(delegate.isEmpty()) {
delegate << e
} else {
def nonZeros = delegate.findAll { it > 0 }
delegate << (nonZeros ? nonZeros.last() + e : e)
}
}
def l = []
l.plusAdd(5)
l.plusAdd(3)
l.plusAdd(2)
assert l == [5, 8, 10]
l = [5, 0]
l.plusAdd(5)
assert l == [5, 0, 10]
l = [1,0]
l.plusAdd(5)
assert l == [1, 0 ,6]
this should be simple and working
incrementingList.add(incrementingList.last() + 5)
check if not empty:
incrementingList.add((incrementingList.size() > 0 ) ? incrementingList.last() + 5 : 5)
,
​def incrementingList = []
incrementingList.add((incrementingList.size() > 0 ) ? incrementingList.last() + 5 : 5)
incrementingList.add((incrementingList.size() > 0 ) ? incrementingList.last() + 5 : 5)
incrementingList.add((incrementingList.size() > 0 ) ? incrementingList.last() + 5 : 5)
println incrementingList​
output:
[5, 10, 15]
Slightly different approach:
def addToList(incList, val) {
if (incList.size() > 0) {
incList << incList.last() + val
} else {
incList << val
}
return incList
}
def incList = []
addToList(incList, 3)
addToList(incList, 2)
addToList(incList, 5)
println incList
output:
[3, 5, 10]
Updated:
def addToList(incList, val) {
if (incList.size() > 0) {
if (val != 0) {
incList << val + incList.last()
} else {
def addVal = incList.size() > 1 ? (incList[-1] - incList[-2]) : incList[-1]
incList << addVal + incList.last()
}
} else {
incList << val
}
}
def incList = []
addToList(incList, 5)
addToList(incList, 3)
addToList(incList, 0)
addToList(incList, 5)
addToList(incList, 0)
println incList
output:
[5, 8, 11, 16, 21]
Suggestion, based on the update request (as I understand it)...
[Edit: the 'empty' case is covered by the 'no non-zero' case, so it can be simplified]
List.metaClass.addAsSum = { e ->
def nonZero = delegate.reverse().find { it != 0 }
if (nonZero != null) {
delegate.add(nonZero + e)
} else {
delegate.add(e)
}
}
Test runs:
def list
// test 0
list = []
list.addAsSum(5)
assert list == [5]
// test 1
list = []
list.addAsSum(5)
list.addAsSum(3)
list.addAsSum(2)
assert list == [5, 8, 10]
// test 2
list = []
list.addAsSum(5)
list.addAsSum(3)
list.addAsSum(0)
list.addAsSum(2)
assert list == [5, 8, 8, 10]
// test 3
list = []
list.addAsSum(0)
list.addAsSum(0)
list.addAsSum(0)
list.addAsSum(2)
list.addAsSum(4)
assert list == [0, 0, 0, 2, 6]
Adding a requirement that on a list filled only with zeros it should simply add the new element, this is my suggestion:
List.metaClass.fill = { n ->
delegate ? delegate.add((delegate.reverse().find { it > 0 } ?: 0) + n) : delegate.add(n)
delegate
}
assert [].fill(1).fill(2).fill(3) == [1, 3, 6]
assert [5, 0].fill(5) == [5, 0, 10]
assert [0, 0].fill(5).fill(5) == [0, 0, 5, 10]
note that I return delegate in the fill() such that you can call it in sequence.

Generating integer partition by its number

I'm trying to generate decent partition of given integer number N numbered K in lexicographical order, e.g. for N = 5, K = 3 we got:
5 = 1 + 1 + 1 + 1 + 1
5 = 1 + 1 + 1 + 2
5 = 1 + 1 + 3
5 = 1 + 2 + 2
5 = 1 + 4
5 = 2 + 3
5 = 5
And the third one is 1 + 1 + 3.
How can I generate this without generating every partition(in C language, but most of all I need algorithm)?
Going to find maximal number in partition(assuming we can find number of partitions d[i][j], where i is number and j is maximal integer in its partition), then decrease the original number and number we are looking for. So yes, I'm trying to use dynamic programming. Still working on code.
This doesn't work at all:
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
FILE *F1, *F2;
main()
{
long long i, j, p, n, k, l, m[102][102];
short c[102];
F1 = fopen("num2part.in", "r");
F2 = fopen ("num2part.out", "w");
n = 0;
fscanf (F1, "%lld %lld", &n, &k);
p = 0;
m[0][0] = 1;
for ( i = 0; i <= n; i++)
{
for (j = 1; j <= i; j++)
{
m[i][j] = m[i - j][j] + m[i][j - 1];
}
for (j = i + 1; j <= n; j++)
{
m[i][j] = m[i][i];
}
}
l = n;
p = n;
j = n;
while (k > 0)
{
while ( k < m[l][j])
{
if (j == 0)
{
while (l > 0)
{
c[p] = 1;
p--;
l--;
}
break;
}
j--;
}
k -=m[l][j];
c[p] = j + 1;
p--;
l -= c[p + 1];
}
//printing answer here, answer is contained in array from c[p] to c[n]
}
Here is some example Python code that generates the partitions:
cache = {}
def p3(n,val=1):
"""Returns number of ascending partitions of n if all values are >= val"""
if n==0:
return 1 # No choice in partitioning
key = n,val
if key in cache:
return cache[key]
# Choose next value x
r = sum(p3(n-x,x) for x in xrange(val,n+1))
cache[key]=r
return r
def ascending_partition(n,k):
"""Generate the k lexicographically ordered partition of n into integer parts"""
P = []
val = 1 # All values must be greater than this
while n:
# Choose the next number
for x in xrange(val,n+1):
count = p3(n-x,x)
if k >= count:
# Keep trying to find the correct digit
k -= count
elif count: # Check that there are some valid positions with this digit
# This must be the correct digit for this location
P.append(x)
n -= x
val = x
break
return P
n=5
for k in range(p3(n)):
print k,ascending_partition(n,k)
It prints:
0 [1, 1, 1, 1, 1]
1 [1, 1, 1, 2]
2 [1, 1, 3]
3 [1, 2, 2]
4 [1, 4]
5 [2, 3]
6 [5]
This can be used to generate an arbitrary partition without generating all the intermediate ones. For example, there are 9253082936723602 partitions of 300.
print ascending_partition(300,10**15)
prints
[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 3, 3, 4, 4, 4, 4, 4, 4, 5, 7, 8, 8, 11, 12, 13, 14, 14, 17, 17, 48, 52]
def _yieldParts(num,lt):
''' It generate a comination set'''
if not num:
yield ()
for i in range(min(num,lt),0,-1):
for parts in _yieldParts(num-i,i):
yield (i,)+parts
def patition(number,kSum,maxIntInTupple):
''' It generates a comination set with sum of kSum is equal to number
maxIntInTupple is for maximum integer can be in tupple'''
for p in _yieldParts(number,maxIntInTupple):
if(len(p) <=kSum):
if(len(p)<kSum):
while len(p) < kSum:
p+=(0,)
print p
patition(40,8,40)
Output:
-------
(40,0,0,0,0,0,0,0)
(39,1,0,0,0,0,0,0)
.
.
.
.

closest to zero [absolute value] sum of consecutive subsequence of a sequence of real values

this is an algorithmic playground for me! I've seen variations of this problem tackling maximum consecutive subsequence but this is another variation as well.
the formal def:
given A[1..n] find i and j so that abs(A[i]+A[i+1]+...+A[j]) is closest to zero among others.
I'm wondering how to get O(n log^2 n), or even O(n log n) solution.
Calculate the cumulative sum.
Sort it.
Find the sequential pair with least difference.
function leastSubsequenceSum(values) {
var n = values.length;
// Store the cumulative sum along with the index.
var sums = [];
sums[0] = { index: 0, sum: 0 };
for (var i = 1; i <= n; i++) {
sums[i] = {
index: i,
sum: sums[i-1].sum + values[i-1]
};
}
// Sort by cumulative sum
sums.sort(function (a, b) {
return a.sum == b.sum ? b.index - a.index : a.sum - b.sum;
});
// Find the sequential pair with the least difference.
var bestI = -1;
var bestDiff = null;
for (var i = 1; i <= n; i++) {
var diff = Math.abs(sums[i-1].sum - sums[i].sum);
if (bestDiff === null || diff < bestDiff) {
bestDiff = diff;
bestI = i;
}
}
// Just to make sure start < stop
var start = sums[bestI-1].index;
var stop = sums[bestI].index;
if (start > stop) {
var tmp = start;
start = stop;
stop = tmp;
}
return [start, stop-1, bestDiff];
}
Examples:
>>> leastSubsequenceSum([10, -5, 3, -4, 11, -4, 12, 20]);
[2, 3, 1]
>>> leastSubsequenceSum([5, 6, -1, -9, -2, 16, 19, 1, -4, 9]);
[0, 4, 1]
>>> leastSubsequenceSum([3, 16, 8, -10, -1, -8, -3, 10, -2, -4]);
[6, 9, 1]
In the first example, [2, 3, 1] means, sum from index 2 to 3 (inclusive), and you get an absolute sum of 1:
[10, -5, 3, -4, 11, -4, 12, 20]
^^^^^

Resources