Mercurial > projects > aid

comparison trunk/ga_maze.d @ 0:4b2e8e8a633e

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Repository setup.
author revcompgeek
date Mon, 03 Mar 2008 19:28:10 -0700
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:4b2e8e8a633e
1 /++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
2 + Genetic Algorithm Maze Test +
3 + Author: Matt P. +
4 + Version: 0.5 +
5 + +
6 + Runs a genetic algorithm test using a maze generation +
7 + and guessing algorithm. +
8 + +
9 + Usage: ga_maze [options] +
10 + Options: +
11 + -h +
12 + Prints this information. +
13 + -v +
14 + Turns on verbose output. +
15 + -m float +
16 + Specifies the mutation rate for each generation. +
17 + Default: 0.05 +
18 + +
19 + -d float +
20 + Specifies the survival rate for each generation. +
21 + Default: 0.5 +
22 + +
23 + -p integer +
24 + Specifies the population for each generation. +
25 + Default: 1000 +
26 + +
27 + -l integer +
28 + Specifies the maximum length of the solution code.+
29 + Default: 20 +
30 + +
31 + -r integer +
32 + Specifies the number of times to repeat the test. +
33 + Default: 20 +
34 + +
35 + -c char +
36 + Specifies the crossover type +
37 + s / 1 - One point crossover +
38 + t / 2 - Two point crossover +
39 + u / 3 - Uniform crossover +
40 + Default: s +
41 + +
42 + -s integer integer +
43 + Specifies the size of the maze +
44 + Default: 7 7 +
45 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++/
46
47
48 module ga_maze;
49
50 import aid.ga;
51 import tango.math.Random;
52 import tango.math.Math;
53 import tango.io.Stdout;
54 //import tango.text.String;
55 import Float = tango.text.convert.Float;
56 import Integer = tango.text.convert.Integer;
57
58 import bcd.sys.times;
59
60 char[] chars = "NESW";
61
62 //int numChars = 20;
63 int stringLen = 28;
64
65 bool[][] maze;
66 int w = 7;
67 int h = 7;
68
69 double average(double[] array){
70 double avg = 0;
71 foreach (double v; array)
72 avg += v;
73 avg /= array.length;
74 return avg;
75 }
76
77 double stdDev(double[] array){
78 double avg = average(array);
79 double stddev = 0;
80 foreach (double v; array){
81 stddev += (v - avg) * (v - avg);
82 }
83 stddev = sqrt(stddev/array.length);
84 return stddev;
85 }
86
87 bool[][] generateMaze(uint width, uint height){
88 bool[][] maze;
89
90 maze.length = width;
91
92 for (auto i = 0; i < width; i++){
93 maze[i].length = height;
94 }
95
96 bool rowStart = true;
97 bool vertical;
98 for (auto i = 1; i < width; i+=2){
99 vertical = rowStart;
100 rowStart = !rowStart;
101
102 for (auto j = 1; j < height; j+=2){
103 maze[i][j] = true;
104 if (vertical)
105 maze[i + ((Random.shared.next(2)-1)?-1:1)][j] = true;
106 else
107 maze[i][j + ((Random.shared.next(2)-1)?-1:1)] = true;
108 vertical = !vertical;
109 }
110 }
111
112 return maze;
113 }
114
115 void printMaze(bool[][] m, bool[][] trail = null,char e = '.', char f = 'X', bool outline = true){
116 if (outline){
117 Stdout("+");
118 for (auto j = 0; j < m[0].length; j++){
119 Stdout("-");
120 }
121 Stdout("+").newline;
122 }
123
124 for (auto i = 0; i < m.length; i++){
125 if (outline) Stdout("|");
126 for (auto j = 0; j < m[i].length; j++){
127 if(m[i][j]) Stdout(""~f);
128 else if(trail && trail[i][j]) Stdout("+");
129 else Stdout(""~e);
130 }
131 if (outline) Stdout("|").newline;
132 }
133 if (outline){
134 Stdout("+");
135 for (auto j = 0; j < m[0].length; j++){
136 Stdout("-");
137 }
138 Stdout("+").newline;
139 }
140 }
141
142 /++++++++++++++++++++++++++ Foo ++++++++++++++++++++++++++++++/
143
144 class Foo {
145 int count = 0;
146 public double calculateFitness(char[] gene){
147
148 bool canMove(int x, int y){
149 return ((x >= 0) && (x < w) && (y >= 0) && (y < h) && (!maze[x][y]));
150 }
151
152 double square(int v){
153 return v*v;
154 }
155
156 /*bool[][] trail;
157
158 trail.length = w;
159 for(uint i = 0; i < w; i++){
160 trail[i].length = h;
161 }*/
162
163 double fitness=0;
164 int x = 0;
165 int y = 0;
166 for (uint i = 0; i < gene.length; i++){
167 //trail[x][y] = true;
168 //Stdout(""~gene[i]);
169 if(gene[i] == 'N' && canMove(x,y-1)) y--;
170 if(gene[i] == 'E' && canMove(x+1,y)) x++;
171 if(gene[i] == 'S' && canMove(x,y+1)) y++;
172 if(gene[i] == 'W' && canMove(x-1,y)) x--;
173
174 if (x == w-1 && y == h-1){
175 return 100;
176 }
177 }
178
179 fitness = 100 - sqrt(square(x-(w-1)) + square(y-(h-1)));
180 /*if(count < 20){
181 Stdout.format("{} {:.4}",gene,fitness).newline;
182 //printMaze(maze,trail);
183 count++;
184 }*/
185 return fitness;
186 }
187
188 public char[] getRandomGenotype(){
189 char[] t;
190 t.length = stringLen;
191 for (int i = 0; i < stringLen; i++){
192 t[i] = chars[Random.shared.next(chars.length)];
193 }
194 return t;
195 }
196
197 public char getRandomChar(uint index){
198 return chars[Random.shared.next(chars.length)];
199 }
200 }
201
202 /++++++++++++++++++++++ End Foo ++++++++++++++++++++++++++++++/
203
204 void usage(){
205 auto s =
206 "Usage: ga_maze [options] \n"
207 "Options: \n"
208 "-h \n"
209 " Prints this information. \n"
210 "-v \n"
211 " Turns on verbose output. \n"
212 "-m float \n"
213 " Specifies the mutation rate for each generation. \n"
214 " Default: 0.05 \n"
215 " \n"
216 "-d float \n"
217 " Specifies the survival rate for each generation. \n"
218 " Default: 0.5 \n"
219 " \n"
220 "-p integer \n"
221 " Specifies the population for each generation. \n"
222 " Default: 1000 \n"
223 " \n"
224 "-l integer \n"
225 " Specifies the maximum length of the solution code.\n"
226 " Default: 20 \n"
227 " \n"
228 "-n integer \n"
229 " Specifies how many possibilities there are for \n"
230 " each character of the code string. \n"
231 " Default: 20 \n"
232 " \n"
233 "-r integer \n"
234 " Specifies the number of times to repeat the test. \n"
235 " Default: 20 \n"
236 " \n"
237 "-c char \n"
238 " Specifies the crossover type \n"
239 " s / 1 - One point crossover \n"
240 " t / 2 - Two point crossover \n"
241 " u / 3 - Uniform crossover \n"
242 " Default: s ";
243 Stdout(s).newline;
244 }
245
246 void main(char[][] args){
247 Foo f = new Foo();
248 GeneticAlgorithm ga = new GeneticAlgorithm();
249 ga.calculateFitness = &f.calculateFitness;
250 ga.getRandomGenotype = &f.getRandomGenotype;
251 ga.getRandomChar = &f.getRandomChar;
252 ga.fitnessThreshold = 100;
253 int rep = 20;
254 bool dotime = true;
255
256
257 for (auto i = 1; i < args.length; i++){
258 char[] arg = args[i];
259 switch (arg){
260 case "-h":
261 usage();
262 return;
263 case "-m":
264 ga.mutationRate = Float.parse(args[++i]);
265 break;
266 case "-d":
267 ga.survivalRate = Float.parse(args[++i]);
268 break;
269 case "-p":
270 ga.startPopulation = Integer.parse(args[++i]);
271 break;
272 case "-l":
273 stringLen = Integer.parse(args[++i]);
274 //ga.fitnessThreshold = stringLen;
275 break;
276 case "-r":
277 rep = Integer.parse(args[++i]);
278 break;
279 case "-v":
280 ga.verbose = true;
281 break;
282 case "-s":
283 w = Integer.parse(args[++i]);
284 h = Integer.parse(args[++i]);
285 stringLen = 2*(w+h);
286 break;
287 case "-c":
288 auto t = args[++i];
289 switch (t){
290 case "s","1":
291 ga.crossoverType = 1;
292 break;
293 case "t","2":
294 ga.crossoverType = 2;
295 break;
296 default:
297 ga.crossoverType = 3;
298 }
299 break;
300 case "-b":
301 ga.bailout = Integer.parse(args[++i]);
302 break;
303 case "-t":
304 dotime = false;
305 break;
306 default:
307 Stdout.format("Unknown parameter: {}",arg).newline;
308 usage();
309 return;
310 }
311 }
312
313 double[] time;
314 double[] gens;
315 long start,end;
316 double generations;
317
318 for(auto i = 0; i < rep; i++){
319 maze=generateMaze(w,h);
320
321 if(dotime) start = iutime();
322 generations = ga.run();
323 if(dotime) end = iutime() - start;
324
325 if(dotime) time ~= end;
326 gens ~= generations;
327 }
328
329 if(dotime) Stdout.format("{0}, {1:.5}, {2}, {3:.5}", average(gens),stdDev(gens),average(time),stdDev(time)).newline;
330 else Stdout.format("{0}, {1:.5}", average(gens),stdDev(gens)).newline; //", {2}, {3:.5}" ,average(time),stdDev(time)
331 }