This map editor will be easier to use compared to the map editor I made for Mungeon. There will be scrollbars to make navigation faster and features such as drawing rectangles and filling in blank tiles. Before the next update, there will be an event editor for handling all of the in-game actions such as animation, special effects, and movement joined onto this editor. Once the main editors are completed, I will move onto finish the battle simulator.
1 Comment
We are trudging towards our goal of completing the following
1. Having a complete story line and major events 2. Designing monsters 3. Designing the battle system 4. Engineering the map editor However, we still need to 1. Create an event editor 2. Create the game engine 3. Design backgrounds and map tiles 4. Program the battle system This is going to be painful. I hope we can at least have a playable 'prototype' before January... Tomorrow marks the start of the 2011-2012 Games Development Club. This year, we have had 28 people sign up to join! We'll have to see how many are actually going to show up...
We already have some ideas regarding what sort of a game we want to make, but first, everyone will be taught the basics of RevMedia (by me). I may post the contents of the lesson material here, if I have the time. If you don't know what RevMedia (known now as LiveCode) is, take a look here: http://www.runrev.com/ I am taking a detour in my attempt at a Go AI. Here is my Monte Carlo method AI for connect four.
The objective of the game is to connect four of your pieces in a straight line, either horizontally, vertically, or diagonally. You can choose up to 7 columns to play your piece in. The piece will then drop to the lowest unoccupied point on the grid. The Monte Carlo method involves thousands of simulations. From a given game position, the program will randomly play it out to completion. For each simulation, it records the first move chosen and the result of the simulation (win or loss). After it has gone through 10000 simulations, it will choose the move with the highest percentage of winning. I was able to beat this AI after losing twice in a row. Considering the fact that I played first on all occasions, the AI is quite strong! いきなり囲碁に挑戦したのは、よくなかったみたいです。そこでConnect-Four というゲームのAIをとりあえず書いてみました。Connect-Four とは自分の駒を四つ繋がるように落としていくゲームです。モンテカルロ法では現在の盤面から乱数を使って終局まで打たせるシミュレーションを何万回やらせます。それによって勝率が一番高かった次の一手を選ぶ訳ですが、著者(先番)は二回負かされたあとやっと一回勝ちました。先手必勝のゲームなのに、このAIなかなかやります。 Here is some crude pseudo-code for the method: void monteCarloMethod() { repeat for i, 0 to 10000 copy_current_game(); next_move = get_random_move(); play_move( next_move) do_simulation(); record_result( next_move, result_of_simulation); end repeat play_move_with_highest_winrate(); } This method alone may not work so well for games that have a significant number of permutations like chess and go. For the last three days, I have been working on a computer program that can play the game of Go, a board game played using black and white stones placed on the intersections of a 19 by 19 board. It may sound impressive, but I barely have anything done at the moment. My program can read and replay kifu (game records) in SGF format. I intend to use the kifu as a starting point for my computer player.
I still need to implement the ko rule, but that shouldn't be a problem. The real problem is the end game scoring (determining dead groups). Does anyone have any good ideas on how to approach this? 過去三日間、囲碁の人工知能プログラムを書いていました。とは言っても、ほとんど何もできてないのですが。 現在 SGF の形式の棋譜なら読み込んで並べられるようになりました。最初は棋譜のデータベースを使って色々やってみるつもりです。 まだコウのルールは作っていません。 ところで問題なのは終局部分の点数計算なのですが、どなたか死石と生き石をうまく判別できる方法をご存知ですか。 ちなみに、これからは日本語と英語のページを一緒にすることにしました。 This sums up what I was able to achieve at the workshop.
The music was made using a database of Bach's Chorales. For every beat of music in the database, the algorithm extracted the pitch and the pitch interval to the next beat and stored as a lexicon. To start the composition, a lexicon is chosen at random. By looking at the pitch interval stored, the algorithm calculates what the next set of pitches should be and finds all the beats with the correct pitches. The process is repeated until no matches are found. I'm currently in UCSC (University of California Santa Cruz) learning how to program a computer to generate music using LISP.
I will post what I came up with after this workshop is finished, which will be in one week. So far, we learned all of the basic functions of LISP and different music generation techniques such as sonification, chaos, cellular automata, Markovian analysis, and rule-based. I will post more on this later. I intend to apply these to the BGMs for my games, including those for GDC. ... and I wasn't there because of sports. Sad, eh?
We didn't finish the Escape game. I am planning on finishing it on C++. We have one more week until the last meet for the Games Development Club. Which means we have one more week to finish our current game project! Most of the graphics are completed. It is now up to our programmer to punch the code in. Let's hope all goes well...
Genre: Puzzle, suspense You are trapped inside a room. The objective is to find a way through the locked door. But you are not alone... Here's what we have so far... 10 minutes of recorded material 1 image 1 graphic of the room 50 lines of code We will be done with this before the 19th of May. |
Adeki Oratuyr
Born 1994. Amateur programmer, game designer, and composer. Head of the Games Development Club. Archives
December 2011
Categories
All
Links
Page Loads
|