Breadth First Search grid shortest path | Graph Theory

Поделиться
HTML-код
  • Опубликовано: 1 дек 2024

Комментарии • 185

  • @jagrit07
    @jagrit07 4 года назад +50

    If you are beginner in graph, watch his videos twice or thrice because then you will feel like no one can explain better than him and you will understand each and every point mentioned in the video. Thank you Sir for this playlist.

  • @hamzahfauzy2150
    @hamzahfauzy2150 6 лет назад +85

    Finally, i'm waiting for this kind of tutorial
    You're great dude

  • @anikethmalyala
    @anikethmalyala 4 года назад +8

    YOU ARE A LIFE SAVER I COULD NEVER FIND A VIDEO LIKE THIS THANK YOU SO MUCH FOR THE GIFT YOUVE BESTOWED UPON THIS LAND

  • @LunchboxdadioMusic
    @LunchboxdadioMusic 3 года назад +4

    I rewrote your pseudocode in JavaScript adjusted it a bit to suit my needs. Worked perfectly right out of the box! Awesome work, William!

    • @saiffyros
      @saiffyros Год назад

      Can you share it?

    • @LunchboxdadioMusic
      @LunchboxdadioMusic Год назад +1

      @@saiffyros
      // ruclips.net/video/KiCBXu4P-2Y/видео.html
      // Globals
      // number of rows and columns
      const R = 10;
      const C = 14;
      // start cell values
      const sr = 4;
      const sc = 0;
      // row and column queue
      const rq = [];
      const cq = [];
      // save the directions found
      const dir = [];
      // variables to track the steps taken
      let move_count = 0;
      let nodes_left_in_layer = 1;
      let nodes_in_next_layer = 0;
      // variable to check if we've reached the end
      let reached_end = false;
      const map = {
      cols: 14,
      rows: 10,
      sSize: 64,
      tsize: 40
      }
      const gameGrid = [
      [ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 9, 0],
      [ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0],
      [ 0, 0, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 1, 0],
      [ 0, 0, 1, 0, 0, 1, 0, 0, 1, 1, 1, 1, 1, 0],
      [ 8, 1, 1, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0],
      [ 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0],
      [ 0, 0, 0, 0, 0, 1, 0, 0, 1, 1, 1, 1, 1, 0],
      [ 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0],
      [ 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 0],
      [ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
      ]
      // R x C matrix of false values aka visited positions
      const visited = [
      [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
      [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
      [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
      [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
      [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
      [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
      [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
      [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
      [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
      [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
      ]
      // north, south, east, west direction vectors
      const dr = [ -1, +1, 0, 0];
      const dc = [ 0, 0, +1, -1];
      function findPath(){
      rq.push(sr);
      cq.push(sc);
      visited[sr][sc] = true;
      // Keep going as long as there are items in the queue
      while(rq.length > 0 || cq.length > 0){
      let r = rq.shift();
      let c = cq.shift();
      if(gameGrid[r][c] === 'E'){
      reached_end = true;
      break;
      }
      check_neighbors( r, c)
      nodes_left_in_layer --;
      if(nodes_left_in_layer === 0){
      nodes_left_in_layer = nodes_in_next_layer;
      nodes_in_next_layer = 0;
      move_count ++;
      }
      }
      if(reached_end){
      return move_count;
      }
      return -1;
      }
      function check_neighbors( r, c){
      for(let i = 0; i < 4; i++){
      let rr = r + dr[i];
      let cc = c + dc[i];
      //skip out of bounds cells
      if(rr < 0 || cc < 0){
      continue;
      }
      if(rr >= R || cc >= C){
      continue;
      }
      // skip blocked or visited cells
      if(visited[rr][cc]){
      continue;
      }
      if(gameGrid[rr][cc] === 0){
      continue;
      }
      let move = [dc[i], dr[i]];
      dir.push(move);
      rq.push(rr);
      cq.push(cc);
      visited[rr][cc] = true;
      nodes_in_next_layer ++;
      }
      }
      findPath();
      function getTile(c, r){
      return gameGrid[r][c];
      }
      export { dir, map, getTile };

  • @WilliamFiset-videos
    @WilliamFiset-videos  4 года назад +73

    A lot of you are asking about how to reconstruct the path, so let me just explain it here.
    To reconstruct the path from the start to the end you need to maintain additional information, namely a 2D matrix which tracks the cell which was used to reach the current cell. Let me call this matrix "prev", short for previous. Every time you advance to the next cell, keep track of which cell you came from in the prev matrix. Once you reach the end of the BFS, start reconstructing the path by beginning at the end node in the prev matrix and work your way to the start node. The obtained path will be in reverse order, so you will need to reverse it.
    This is explained in more detail in the BFS video except that the prev matrix is 1D for the general case: ruclips.net/video/oDqjPvD54Ss/видео.html

    • @austind649
      @austind649 4 года назад

      Hi William - @14:47, shouldn't we enqueue prior to "continue" on the # equality check? Otherwise certain positions would never be processed.

    • @marjank7781
      @marjank7781 4 года назад +1

      @@austind649 Hi Austin, whenever we reach a '#' cell, that cell can not be visited as it's an obstacle so can not be in our path. so we don't need to process it and we need to find another path through a reachable cell. The whole point here is only go through the path that is reachable.

    • @ss4036
      @ss4036 4 года назад +1

      Can you please share the code for saving the actual path along with the moves?

    • @siddharthmagadum16
      @siddharthmagadum16 4 года назад

      @@ss4036 Here is my code in C++, implemented by the williamfiset approach, which stores the the path and also shows BFS path taken in the matrix.
      pastebin.com/TeMDhspF

    • @Steamed111
      @Steamed111 4 года назад

      Hi, what about if you can move diagonally? Mine calculates the distance correctly and I had the path reconstruction working for the cardinal directions but once I add diagonal vectors then the path reconstruction works strangely. I know its probably hard to say without looking at my code but how do you modify the above algorithm to account for diagonal? Or should it "just work" if you add the new vectors?

  • @shivamsahni6720
    @shivamsahni6720 4 года назад +3

    William your explanation and animation creates a clear picture in my head about the underlying Concept, Thanks alot man!

  • @JustMe111094
    @JustMe111094 5 лет назад +17

    This channel is a gem. very high quality content in these tutorial videos!

  • @brunokawka
    @brunokawka 4 года назад +2

    This is probably one of the most helpful 'Algorithms & Data Structures' channels I've yet encountered. The idea of maintaining multiple queues for each dimension is really handy.

    • @alikhansmt
      @alikhansmt 4 года назад +1

      Can you please explain why its better to have separate queues for each dimension? I cant see how its better than having one queue with the dimensions encapsulated into one object..

    • @brunokawka
      @brunokawka 4 года назад +2

      @@alikhansmt It's just a personal preference of mine. I've found this approach quicker to implement, and reuse over larger code structures, where I don't like to maintain a vast amount of structs, dot callbacks etc... Speaking in terms of competitive programming of course.

    • @alikhansmt
      @alikhansmt 4 года назад

      @@brunokawka I see thanks!

  • @create_space812
    @create_space812 3 года назад +2

    Thank you William for the great video. Fellow Canadian here. thank you for the high-quality content!

  • @JinkProject
    @JinkProject 5 лет назад +2

    i love how slow you speak and how thoroughly you explain everything. thanks!

  • @harishshankar682
    @harishshankar682 4 года назад +2

    Best explanation for coding problems and approaches!! Thanks for making such high quality content on RUclips

  • @hemanthchalla4833
    @hemanthchalla4833 5 лет назад +2

    I was hammering about how to solve these ki9nd of problems. You are a savior! Please make these kind of tutorials more. I cannot thank you enough. Subscribed!

  • @Mnkmnkmnk
    @Mnkmnkmnk 3 года назад +18

    An alternative to using 'nodes_left_in_layer' and 'nodes_in_next_layer' variables is to store the distance/level along with the coordinates into your queue.
    It does have the downside of using an Object or Array to hold the coordinates and the level, but is easier to understand.

    • @rajns8643
      @rajns8643 Год назад

      Thanks for this approach! Very intuitive indeed!

  • @rohitkumar-rq6qh
    @rohitkumar-rq6qh 6 лет назад +10

    This is nice.I have seen this trick used by competitive programmers for traversing in the grid.

  • @andrewstien7179
    @andrewstien7179 3 года назад +4

    What a great tutorial! A great explanation complemented by incredibly clear slides and animations. Thanks.

  • @francaniilista
    @francaniilista Месяц назад

    Thank you so much, William, tons of good information in a very nice format

  • @sarthakshah1058
    @sarthakshah1058 6 лет назад +40

    Your videos are amazing, but there is one possible improvement:
    run through the pseudo-code with the animation for the algorithm, so visualizing the pseudo-code becomes easier

    • @adityasingh8973
      @adityasingh8973 4 года назад +1

      Video for visualization!!
      Watch till end for better visualization.
      Used exactly same concept as explained!!
      ruclips.net/video/EdJa84ymIXg/видео.html

  • @Sumit-sl5lp
    @Sumit-sl5lp 5 лет назад +3

    this is exactly what i needed to know to solve graph problems, thanks for the video :) watching your tutorial for the first time, should have watched it earlier.

  • @tc07client5
    @tc07client5 6 лет назад +2

    I am preparing for IOI and this video was very Helpful, continue to make great contents like this. Thanks!

  • @Toasterparty
    @Toasterparty 4 года назад

    My mind is blown at how many things can be modeled as a graph. I go out into the world and just see graphs now.

  • @dj1984x
    @dj1984x Год назад

    excellent explanation. I particularly liked your concept for using direction vectors

  • @vaibhavlodha5398
    @vaibhavlodha5398 5 лет назад +2

    Great video. thank you very much. This is exactly what I needed to understand how to solve graph problem using adjacency matrix. Thank you once again!

  • @josephwong2832
    @josephwong2832 4 года назад +2

    Thanks a lot Will! This is from cracking the coding interview and your walkthrough is well explained. I'm subbing to your channel now

  • @Hajjat
    @Hajjat 5 лет назад +3

    Amazing explanation, thanks for the video! I've been always avoiding BFS and going with DFS since I wasn't comfortable with it, but not anymore! Subscribed to your channel... Will be waiting for more tutorials from you!

  • @adambruce4284
    @adambruce4284 2 года назад +1

    This is some of the best fucking content I've ever seen. Absolutely fucking incredible. 10/10 would peek again!

  • @chingizmardanov
    @chingizmardanov 6 лет назад +1

    You are just brilliant man! I hope your channel grows big!

  • @user-or7ji5hv8y
    @user-or7ji5hv8y 5 лет назад

    really like the summary in the end, as well. it's a nice touch.

  • @huseyinbarin1653
    @huseyinbarin1653 2 года назад +2

    Very clear approach! Just would like to share something regarding using 2 queues for 2D Matrixes, it just affects the performance very badly single Queue implementation is much more efficient. Other than this, great explanation thanks.

    • @rajns8643
      @rajns8643 Год назад +1

      Efficient in terms of memory complexity, right?
      (Because handling so many queues in n-dimensional space would require a lot of memory I presume).

  • @zinda__bad
    @zinda__bad 5 месяцев назад

    Brilliant video, thank you so much for these.
    I think you could also modify solve() to return a prev collection as in the BFS Shortest Path video, and add the end cell coordinates or marker character ('E') as a parameter to the enclosing function. Then you no longer need all the global counting variables (nodes_left_in_layer etc). You would need to perform [number of steps in shortes path] iterations over prev to get the shortest path and return shortest_path.size() - 1.
    Maybe not optimal since you have to iterate over the shortest path too, but as long as you still break inside solve() when reached_end = True, the time complexity should stay the same.

  • @ego_sum_liberi
    @ego_sum_liberi 6 лет назад +5

    Bravo! you are one of a kind, keep rocking!!!

  • @srv7151
    @srv7151 4 года назад

    Best video on 2d Grid Thank You for the explanation

  • @gayanhewa
    @gayanhewa 5 лет назад +1

    Good stuff, for my dumb brain it wasn't clear in the first glance why we needed the nodes_in_next_layer and nodes_left_in_layer and how it's used. Had to replay and follow through the algorithm to understand its use.

    • @eurekagao2078
      @eurekagao2078 5 лет назад +12

      Try to think this way: BFS is visiting graph layer by layer. Initially the nodes_left_in_layer = 1; supposed that there're 2 adjacent nodes (A, B)connecting the starting node(S), so the nodes_in_next_layer will be incremented to 2 after the explorer_neighbours(), then the nodes_left_in_layer is decremented to 0. Here, we reset the nodes_left_in_layer to 2 and nodes_in_next_layer to 0 and increment the move_count to 1. It means after visiting layer 1, we are expecting to visit 2 nodes in layer 2. We can image node A has 3 adjacent nodes (C, D, E) and node B has 1 adjacent nodes (F). Now try to go-thru layer 2 with the imaginary graph - we visit A, nodes_in_next_layer is set to 3, nodes_left_in_layer is set to 1, then visit B, nodes_in_next_layer is set to 4 and nodes_left_in_layer to 0. Here, we are finishing visit layer 2 and expecting to visit layer 3 with 4 coming nodes.
      snap shots of queue:
      (S)
      (A) (B) // nodes_in_next_layer = 2, after visiting layer 1;
      (B) (C) (D) (E) // pop (A) and push all neighbours (C) (D) (E)
      (C) (D) (E) (F) // pop (B) and push neighbour (E)
      // nodes_in_next_layer = 4, after visiting layer 2;
      Hope this helps.

    • @jsarvesh
      @jsarvesh 3 года назад +3

      i think level would be better name compared to layer as BFS processes nodes which are at same level/distance from current node. Once we process all the nodes in current level, we move onto the next level. BFS is also called commonly referred to as level order traversal

    • @hil449
      @hil449 2 года назад +1

      @@jsarvesh yea, level is a much better name. I also find easier to code and understand passing an additional information into the object/struct of the queue, that is the level so when i push neighbors to the queue i increment this level/distance like this:
      q.push((position){.level=nextLevel, .row=nextRow, .col=nextCol, .distance=pos.distance + 1});

  • @dmitriy5763
    @dmitriy5763 3 года назад +1

    Thank you so much! You definitely have talent in explaining!

  • @hagergamal2856
    @hagergamal2856 4 года назад +1

    This is amazing ♥️
    Please keep solving that kind of problems ♥️♥️😍

  • @aguluman
    @aguluman 7 месяцев назад

    A fellow student from hyperskill, psted your video link. On the lee algorithm course conent.
    It is helpful.

  • @anikethmalyala
    @anikethmalyala 4 года назад

    This helped me so much you don't even know TYSM

  • @senthilkumar5
    @senthilkumar5 5 лет назад +1

    Thank you so much for simplifying it

  • @ZzBiazZ
    @ZzBiazZ Год назад

    Awesome video, thank you so much!

  • @Nealpa
    @Nealpa 2 года назад

    Such a great explanation!!

  • @nurlan1666
    @nurlan1666 2 года назад +1

    Спасибо вам, за хорошее объяснение

  • @mohdfayaq3037
    @mohdfayaq3037 5 лет назад

    Great buddy! Got it in the first go!

  • @germin1983
    @germin1983 2 года назад +1

    Now that I know the shortest path in value. How do i highlight the path it took.

  • @codeblooded6760
    @codeblooded6760 4 года назад +5

    I didnt understand how r u calculating nodes left in layer and nodes in next layer

  • @samtux762
    @samtux762 5 лет назад

    I also use two int arrays instead of an array of int,int objects. Before I was shy to tell about it because it is "not OOP". Now my design is backed by this video.

  • @TheMato1112
    @TheMato1112 4 года назад

    Thank you very much. It helped me a lot.

  • @grandson_f_phixis9480
    @grandson_f_phixis9480 6 месяцев назад

    Thank you very much sir!!

  • @andreamengoli4656
    @andreamengoli4656 5 лет назад +1

    Thank you.. this is helpful for making games as well

  • @InfinityFnatic
    @InfinityFnatic 5 лет назад +1

    Amazing tutorial! Keep filming please :)

  • @sairamankilambi5007
    @sairamankilambi5007 4 года назад

    Quality content. Really appreciate it👍

  • @brayaon
    @brayaon 5 лет назад

    thank you so much for these videos!

  • @tejasri2564
    @tejasri2564 4 года назад +2

    what is the time complexity for this?

  • @dudeov
    @dudeov 4 года назад

    Amazing video! Thanks a ton!

  • @juliesz78
    @juliesz78 4 года назад

    Awesome video!!! Thanks bro!

  • @AsliArtistVlogs
    @AsliArtistVlogs 3 года назад +1

    What do you mean by a 'layer'? (I am sorry if I missed it but I think its missing from the video)

  • @kompeterPC
    @kompeterPC 5 лет назад

    Good explanation, thanks a lot

  • @Garentei
    @Garentei 4 года назад +1

    Thanks for the multiple queue trick. Storing them as pairs does not scale well to higher dimensions and p.first and p.second looks really ugly and ambiguous.

  • @sumant9120
    @sumant9120 4 года назад +1

    Could anyone explain what is the use of 'nodes_left_in_layer' and 'nodes_in_next_layer' variables?

  • @catalinadascalete7094
    @catalinadascalete7094 5 лет назад

    Great tutorial! Congrats

  • @prakhardoshi6525
    @prakhardoshi6525 5 лет назад +1

    Brilliant ! Thank you :)

  • @goestriyadi
    @goestriyadi Год назад

    I created a ArrayList variable to store the steps taken, but it store all the block the BFS explored instead of the green block that shown in the video.
    How can I only store the shortest block need to be explored only?

  • @LawZist
    @LawZist 6 лет назад +3

    keep with the good job!

  • @ayoubed3496
    @ayoubed3496 5 лет назад +15

    Your videos are amazing, really!
    Can I know where do you run your animations, please?

  • @mukulbakshi28
    @mukulbakshi28 4 года назад

    Man, wish I could see this before. Encountered this same problem in my First Round of Google Interview For Software Engineer.

    • @kickhuggy
      @kickhuggy 4 года назад

      Dude, their on site is all day variations of BFS problems, there's no hope lol

    • @hil449
      @hil449 2 года назад

      @@kickhuggy really? thats actually pretty cool lol

  • @a2pha
    @a2pha 2 года назад

    8:36 I understand the "Fill" method to get all the tiles, yet how are you getting this path from that ?

  • @PavLosMavRis
    @PavLosMavRis 4 года назад

    What about the Robot Vacuum Cleaner problem, a double array 3x3 a robot is somewhere in the grid and in some cells there is dirt and the robot must find the optimal path in order to clean the cells and doing that with the BFS algorithm. Any suggestions?

  • @adityamaurya3673
    @adityamaurya3673 3 года назад +1

    what happen if we are allowed to go in 8 direction diagonal also included then i think only BFS will not work anybody??

  • @LunaMarlowe327
    @LunaMarlowe327 5 лет назад

    This guy is insane

  • @kolick7547
    @kolick7547 3 года назад

    How is it possible to get returned the coords for the path? I dont think the queue in the end contains the path or my implementation is wrong.

  • @ujjwalrawat5105
    @ujjwalrawat5105 2 года назад

    I don't understand the nodes left in layer and nodes in next layer part. Can someone link to a C++ solution to this problem using STL?

  • @anonymoussloth6687
    @anonymoussloth6687 4 года назад

    for the dungeon problem, why do we have to use bfs? why not dfs?

  • @journeytowardslife9830
    @journeytowardslife9830 3 года назад

    How do we know that this algorithm gives the shortest path ?

  • @vanyastaleva415
    @vanyastaleva415 5 лет назад

    Am I the only one who missed the part where we select the path that he colored in green? What's explained in this video only shows how many steps we need to reach the end, but not the actual path.

  • @_va3y
    @_va3y 2 года назад

    is there a related leetcode problem to practice?

  • @chamathtoo
    @chamathtoo 7 месяцев назад

    thanks it s really great bt how can I get the source code for this?

  • @icecreamicequeen28
    @icecreamicequeen28 4 года назад

    THANK YOU!

  • @darshanabawane291
    @darshanabawane291 3 года назад

    I have one question about these can you help me please

  • @sameernema3139
    @sameernema3139 4 года назад

    hey could we get the complete source code file?

  • @davidfitzsimons2535
    @davidfitzsimons2535 3 года назад

    how do you show the path once its done?

  • @THECMON6969
    @THECMON6969 4 года назад

    11:16 does this pseudo code applies for a 3d grid graph? im confused you mention to use 3 queues for 3ds grid graph and then went straight back to 2d pseudo code

  • @rmp251
    @rmp251 2 года назад

    8:55 how do we find the number of steps without tracing the path itself?

  • @muhammadsaifuddinjazulijaz3898
    @muhammadsaifuddinjazulijaz3898 4 года назад

    Your videos are amazing, but how do I get the short path if reached end ?, because my move_count data has already been calculated

  • @tempregex8520
    @tempregex8520 3 года назад

    can someone please help me understand, why is it we choose BFS over DFS to find "shortest path" ? i believe it has to be the way BFS traverses a graph as compared to DFS, but looking for a better explanation

  • @sonluuh
    @sonluuh 3 года назад

    Hi, I have a (maybe so stupid question). BFS cannot use as a shortest path finding with weighted graph right? (it just work with unweight - or all the edge have the same priority). I am not sure about this because at the video Overview, you said it can be use as shortest path finding algorithms. (But to my knowlege, it is impossible). Maybe i got mistake, thanks.

  • @huyvuquang2041
    @huyvuquang2041 4 года назад

    Big like for you. Thanks a lot :))

  • @Dat2Phit
    @Dat2Phit 4 года назад

    4:30 Send this to somebody without context.

  • @urbantech28
    @urbantech28 4 года назад +1

    Used queue and showing Stack ... :(

  • @priyanshuraturi8808
    @priyanshuraturi8808 4 года назад

    what if we are allowed to take off one obstacle???

  • @Sumit-sl5lp
    @Sumit-sl5lp 5 лет назад

    Hi William, shouldn't the move count be move_count+1(if reached_end is true) as we are breaking the loop and not incrementing move_count for the last layer(where end node was found)?

    • @prokiddie3520
      @prokiddie3520 5 лет назад

      No. When reaching the layer containing the E, the step to get to that layer is already accounted for.

  • @tilakramyadav559
    @tilakramyadav559 4 года назад

    Thanks a lot

  • @muzafferckay2609
    @muzafferckay2609 4 года назад

    You are the best

  • @saadahmad485
    @saadahmad485 3 года назад

    Damn this was good

  • @AnkitSharma-gf6ok
    @AnkitSharma-gf6ok 4 года назад

    thanks ,,, that was helpfull

  • @abhishekp4955
    @abhishekp4955 6 лет назад

    Can you please share the link for the code, not able to find in your github link

    • @WilliamFiset-videos
      @WilliamFiset-videos  6 лет назад +1

      I think there's only the pseudo code I show in this video. The previous video explaining generalized BFS has code though :)

  • @kartikeyasrivastava987
    @kartikeyasrivastava987 4 года назад

    Are you the same guy from freecodecamp video?

  • @dkfitness825
    @dkfitness825 6 лет назад

    may i have a code plz

  • @akibmaredia9911
    @akibmaredia9911 5 лет назад

    where can if find the video about tracing back the path?

    • @afrahhassan741
      @afrahhassan741 4 года назад

      I guess you're talking about backtracking technique which is used in the DFS method of maze solving

  • @zinekhalaf2
    @zinekhalaf2 Год назад

    انت وحش حقيقي

  • @geraldbustos
    @geraldbustos 4 года назад

    Notable

  • @shivakumar-gr4go
    @shivakumar-gr4go 6 лет назад

    Looking at the order of direction vectors which is (-1,0), (+1,0), (0,+1), (0, -1)...it looks like you are interested to move row wise first and then column wise. Then how does it become breadth first algorithm. As BFA says, we to move across the row first right.

    • @WilliamFiset-videos
      @WilliamFiset-videos  6 лет назад +1

      Hi Shiva, it doesn't matter which direction vector is applied first because you're doing the BFS layer by layer and adding newly discovered nodes to the end of the queue.

    • @shivakumar-gr4go
      @shivakumar-gr4go 6 лет назад

      That's exactly is my question.. how does this approach taking breadth first. With the use of vectors we go on finding the next node to traverse and looking at the these vectors, traversing need not be breadth wise. Because you are looking for next node above and below the current node first. I hope it makes sense.

    • @WilliamFiset-videos
      @WilliamFiset-videos  6 лет назад +2

      Let me try and make this as clear as possible. For any given node, we add all the neighbors of that node to the list of nodes that need to be visited (unless the neighboring node has already been visited, is not traversable or something weird). The list of nodes to be visited is stored in a queue data structure. The queue has the property that the most recent node added to the queue is found at the end and the oldest node which has been in the queue the longest at the beginning. This means that we can add nodes to the queue and know that they will eventually get processed at a later stage. When the BFS starts you add the starting node to the queue to indicate that it should be visited. The algorithm begins by dequeuing (removing) the start node from the beginning of the queue, after which you add the left, right, down and up neighbors of the start node to the queue in that order (which shouldn't matter). Then the starting node's left node is first in the queue, so you add its left, right, down and up neighbors to the queue, then the starting node's right node if first in the queue, so you add its left, right, down and up neighbors to the queue, and etc... So the algorithm circles around if you will and the graph is explored layer by layer

    • @shivakumar-gr4go
      @shivakumar-gr4go 6 лет назад

      Thanks for the detailed explanation ....

  • @orocimarosay1447
    @orocimarosay1447 3 года назад +1

    But it is better to keep everything in one queue due to cache efficiency and you could just use a struct Point{int x; int y;};