Initial file setup for project

CMakeLists.txt

@@ -1,7 +1,7 @@
 cmake_minimum_required(VERSION 3.10)
 
 project(
-    n-queens
+    n_queens
     LANGUAGES CXX)
 
 set(CMAKE_CXX_STANDARD 11 CACHE STRING "The C++ standard to use")

src/CMakeLists.txt

@@ -1,6 +1,8 @@
-add_executable(n-queens
+add_executable(n_queens
-    ./n-queens.cpp
+    ./n_queens.cpp
+    ./chess.cpp
+    ./genetic_algorithm.cpp
 )
 
-target_include_directories(n-queens PUBLIC ${CMAKE_CURRENT_LIST_DIR})
+target_include_directories(n_queens PUBLIC ${CMAKE_CURRENT_LIST_DIR})
 

src/genetic_algorithm.cpp

@@ -0,0 +1,145 @@
+#include "genetic_algorithm.hpp"
+#include <stdio.h>
+#include <stdlib.h>
+#include <time.h>
+
+/* Prints the population strings in a line */
+void print_population(unsigned char population_in[]) {
+    int iterator = 0;
+    int member_count = 0;
+    char cur_member;
+    while (member_count < N){
+        cur_member = population_in[member_count];
+        while (iterator < L){
+            if (cur_member & 0x80){
+                printf("1");
+            }
+            else{
+                printf("0");
+            }
+            cur_member = cur_member << 1;
+            iterator++;
+        }
+        member_count++;
+        iterator = 0;
+        printf(" ");
+    }
+    printf("\n");
+}
+
+/* Fitness is determined by the number of 
+   1's in the bitstring. */
+int get_fitness(unsigned char string_in){
+    int count = 0;
+    unsigned char temp = string_in;
+    while (temp){
+        if (temp & 0x01){
+            count++;
+        }
+        temp = temp >> 1;
+    }
+    return count;
+}
+
+/* Randomly initialize the first population */
+void init_population(unsigned char* population){
+    int i;
+    for ( i=0; i<N; i++){
+        population[i] = (char)(rand() % 0xFF);
+    }
+}
+
+
+/* Perform selection of population members based on fitness */
+void do_selection(unsigned char* population, int* selected){
+    int fitness[N] = {0};
+    int fitness_sum = 0;
+    int i, j;
+    // get fitness for all members of population
+    for ( i=0; i<N; i++ ){
+        fitness[i] = get_fitness(population[i]);
+        fitness_sum += fitness[i];
+    }
+
+
+    // this is simple fitness proportional selection
+    // (roulette wheel sampling)
+    int roll;
+    int temp_sum = 0;
+    int selection;
+    for ( i=0; i<N; i++ ){
+        temp_sum = 0;
+        roll = rand()%fitness_sum;
+        for ( j=0; j<N; j++ ){
+            temp_sum += fitness[j];
+            if ( roll < temp_sum ){
+                selection = j;
+                break;
+            }
+        }
+        selected[i] = selection;
+    }
+}
+
+
+/* compute a mask to use when crossing over parents*/
+unsigned char get_mask(int locus_in){
+    int i = 0;
+    unsigned char ret;
+    for( i=0; i<locus_in; i++ ){
+        ret = ret << 1;
+        ret ^= 0x01;
+    }
+    return ret;
+}
+
+/* crossover members with probability P_c
+   if no crossover, then clone parents   */
+void do_crossover(unsigned char* population, int* selected){
+    double crossover_roll;
+    int crossover_locus;
+    int i;
+    unsigned char temp1;
+    unsigned char temp2;
+    unsigned char mask;
+    unsigned char temp_population[N];
+
+    for ( i=0; i<N; i+=2){ 
+        crossover_roll = ((double)rand())/((double)RAND_MAX);
+        temp1 = 0;
+        temp2 = 0;
+        if(crossover_roll <= P_c){  //crossover
+            crossover_locus = rand()%L;
+            mask = get_mask(crossover_locus);
+            temp1 = population[selected[i]] & mask;
+            temp1 ^= population[selected[i+1]] & ~mask;
+            temp2 = population[selected[i+1]] & mask;
+            temp2 ^= population[selected[i]] & ~mask;
+            temp_population[i] = temp1;
+            temp_population[i+1] = temp2;
+        }
+        else{  //clone
+            temp_population[i] = population[selected[i]];
+            temp_population[i+1] = population[selected[i+1]];
+        }
+    }
+
+    //copy back to population
+    for ( i=0; i<N; i++ ){
+        population[i] = temp_population[i];
+    }
+}
+
+
+void do_mutation(unsigned char* population){
+    double mutation_roll;
+    int i, j;
+    for ( i=0; i<N; i++){
+        for ( j=0; j<L; j++ ){
+            mutation_roll = ((double)rand())/((double)RAND_MAX);
+            if ( mutation_roll <= P_m ){
+                population[i] ^= (1<<j);   //toggle bit
+            }
+        }
+    }
+}

src/genetic_algorithm.hpp

@@ -0,0 +1,19 @@
+#ifndef GENETIC_ALGORITHM_HPP
+#define GENETIC_ALGORITHM_HPP
+
+#define P_c 0.7     //crossover probability (typical val.)
+#define P_m 0.001   //mutation probability  (typical val.)
+#define N 8         //population size (change to something even)
+#define L 8         //string length (don't change)
+#define G 10000       //number of generations (something huge)
+
+void print_population(unsigned char population_in[]);
+int get_fitness(unsigned char string_in);
+void init_population(unsigned char* population);
+void do_selection(unsigned char* population, int* selected);
+unsigned char get_mask(int locus_in);
+void do_crossover(unsigned char* population, int* selected);
+void do_mutation(unsigned char* population);
+
+#endif // GENETIC_ALGORITHM_HPP
+

src/n_queens.cpp

@@ -0,0 +1,35 @@
+#include "chess.hpp"
+#include "genetic_algorithm.hpp"
+#include <stdio.h>
+#include <stdlib.h>
+#include <time.h>
+
+int main(){
+    unsigned char population[N] = {0};
+    int selected[N] = {-1};
+    int generation_count = 0;
+    int i;
+    srand(time(NULL));
+
+
+    //basic genetic algorithm skeleton
+    init_population(population);
+    print_population(population);
+
+    while (generation_count < G) {
+        do_selection(population, selected);
+        do_crossover(population, selected);
+        do_mutation(population);
+        printf("%4d: ", generation_count);
+        print_population(population);
+        for (i = 0; i < N; i++) {
+            if (population[i] == 0xFF) {
+                printf("Max fit reached.\n");
+                return 0;
+            }
+        }
+        generation_count++;
+    }
+
+    return 0;
+}