feat: plot_data.py to compare results from paper and current

DiscreteENN_TSP_table.csv

@@ -1,18 +1,18 @@
-name points error time(us) time per city(us) distance optimal_distance
-berlin52,52,1.580000,255.000000,4.900000,7661.009766,7542.000000
-st70,70,2.170000,661.000000,9.440000,689.640015,675.000000
-eil76,76,4.850000,741.000000,9.750000,564.099976,538.000000
-pr76,76,2.490000,1051.000000,13.830000,110850.609375,108159.000000
-kroA100,100,3.750000,1695.000000,16.950001,22080.689453,21282.000000
-kroC100,100,3.170000,1591.000000,15.910000,21405.820312,20749.000000
-kroD100,100,6.780000,1609.000000,16.090000,22737.619141,21294.000000
-rd100,100,8.360000,1675.000000,16.750000,8571.240234,7910.000000
-eil101,101,6.060000,2053.000000,20.330000,667.119995,629.000000
-lin105,105,6.090000,1564.000000,14.900000,15254.559570,14379.000000
-ch130,130,7.320000,3135.000000,24.120001,6557.279785,6110.000000
-ch150,150,4.270000,6040.000000,40.270000,6806.850098,6528.000000
-tsp225,225,4.060000,18750.000000,83.330002,4074.899902,3916.000000
-a280,280,7.400000,32272.000000,115.260002,2769.820068,2579.000000
-pcb442,442,10.520000,137540.984375,311.179993,56120.769531,50778.000000
-pr1002,1002,7.580000,1523460.000000,1520.420044,278676.968750,259045.000000
-pr2392,2392,10.000000,32860496.000000,13737.669922,415825.000000,378032.000000
+name,points,error,time(us),time per city(us),distance,optimal_distance
+berlin52,52,5.720000,250.000000,4.810000,7973.629883,7542.000000
+st70,70,4.030000,494.000000,7.060000,702.210022,675.000000
+eil76,76,6.100000,755.000000,9.930000,570.799988,538.000000
+pr76,76,4.440000,765.000000,10.070000,112965.250000,108159.000000
+kroA100,100,3.480000,1576.000000,15.760000,22022.210938,21282.000000
+kroC100,100,4.320000,1349.000000,13.490000,21645.830078,20749.000000
+kroD100,100,3.870000,2084.000000,20.840000,22118.060547,21294.000000
+rd100,100,2.960000,1385.000000,13.850000,8143.919922,7910.000000
+eil101,101,5.490000,1976.000000,19.559999,663.530029,629.000000
+lin105,105,5.900000,1765.000000,16.809999,15227.559570,14379.000000
+ch130,130,5.600000,3071.000000,23.620001,6451.919922,6110.000000
+ch150,150,9.550000,5407.000000,36.049999,7151.419922,6528.000000
+tsp225,225,4.660000,19827.000000,88.120003,4098.509766,3916.000000
+a280,280,7.160000,36695.000000,131.050003,2763.530029,2579.000000
+pcb442,442,7.860000,132595.000000,299.989990,54771.109375,50778.000000
+pr1002,1002,8.100000,1413649.000000,1410.829956,280030.656250,259045.000000
+pr2392,2392,10.050000,22980268.000000,9607.139648,416033.687500,378032.000000

DiscreteENN_TSP_table.txt

@@ -1,18 +1,18 @@
 name points error time(us) time per city(us) distance optimal_distance
-berlin52 52 1.580000 255.000000 4.900000 7661.009766 7542.000000
-st70 70 2.170000 661.000000 9.440000 689.640015 675.000000
-eil76 76 4.850000 741.000000 9.750000 564.099976 538.000000
-pr76 76 2.490000 1051.000000 13.830000 110850.609375 108159.000000
-kroA100 100 3.750000 1695.000000 16.950001 22080.689453 21282.000000
-kroC100 100 3.170000 1591.000000 15.910000 21405.820312 20749.000000
-kroD100 100 6.780000 1609.000000 16.090000 22737.619141 21294.000000
-rd100 100 8.360000 1675.000000 16.750000 8571.240234 7910.000000
-eil101 101 6.060000 2053.000000 20.330000 667.119995 629.000000
-lin105 105 6.090000 1564.000000 14.900000 15254.559570 14379.000000
-ch130 130 7.320000 3135.000000 24.120001 6557.279785 6110.000000
-ch150 150 4.270000 6040.000000 40.270000 6806.850098 6528.000000
-tsp225 225 4.060000 18750.000000 83.330002 4074.899902 3916.000000
-a280 280 7.400000 32272.000000 115.260002 2769.820068 2579.000000
-pcb442 442 10.520000 137540.984375 311.179993 56120.769531 50778.000000
-pr1002 1002 7.580000 1523460.000000 1520.420044 278676.968750 259045.000000
-pr2392 2392 10.000000 32860496.000000 13737.669922 415825.000000 378032.000000
+berlin52 52 5.720000 250.000000 4.810000 7973.629883 7542.000000
+st70 70 4.030000 494.000000 7.060000 702.210022 675.000000
+eil76 76 6.100000 755.000000 9.930000 570.799988 538.000000
+pr76 76 4.440000 765.000000 10.070000 112965.250000 108159.000000
+kroA100 100 3.480000 1576.000000 15.760000 22022.210938 21282.000000
+kroC100 100 4.320000 1349.000000 13.490000 21645.830078 20749.000000
+kroD100 100 3.870000 2084.000000 20.840000 22118.060547 21294.000000
+rd100 100 2.960000 1385.000000 13.850000 8143.919922 7910.000000
+eil101 101 5.490000 1976.000000 19.559999 663.530029 629.000000
+lin105 105 5.900000 1765.000000 16.809999 15227.559570 14379.000000
+ch130 130 5.600000 3071.000000 23.620001 6451.919922 6110.000000
+ch150 150 9.550000 5407.000000 36.049999 7151.419922 6528.000000
+tsp225 225 4.660000 19827.000000 88.120003 4098.509766 3916.000000
+a280 280 7.160000 36695.000000 131.050003 2763.530029 2579.000000
+pcb442 442 7.860000 132595.000000 299.989990 54771.109375 50778.000000
+pr1002 1002 8.100000 1413649.000000 1410.829956 280030.656250 259045.000000
+pr2392 2392 10.050000 22980268.000000 9607.139648 416033.687500 378032.000000

main.cpp

@@ -173,7 +173,7 @@ int main(int argc, char** argv)
  }
  table_file.close();
  std::ofstream csv_file{ "DiscreteENN_TSP_table.csv" };
- csv_file << table_header << std::endl;
+ csv_file << utils::subtituteStr(table_header, "\t", ",") << std::endl;
  for (auto it{ optimal_infos.begin() }; it != optimal_infos.end(); ++it) {
  TSPInfo& opt_info{ *it };
  int info_pos{ utils::vectFind(opt_info, infos) };

plot_data.py

@@ -1,36 +1,157 @@
 import numpy as np
+import math
 import os
 from os import path as osPath
 from matplotlib import pyplot as plt
+import pandas as pd
 
 DATA_DIR = os.getcwd()
-DATA_FILENAME = "./enn_tsp_data.csv"
+DATA_FILENAME1 = "./enn_tsp_data.csv"
+DATA_FILENAME2 = "./DiscreteENN_TSP_table.csv"
 
-DATA_FILE = osPath.join(DATA_DIR, DATA_FILENAME)
+DATA_FILE1 = osPath.join(DATA_DIR, DATA_FILENAME1)
+DATA_FILE2 = osPath.join(DATA_DIR, DATA_FILENAME2)
 
-def plotData(data, load_data: bool, sequence: bool):
- data = np.loadtxt(data, delimiter=',') if load_data else data
- # data = np.genfromtxt(data, delimiter=',') if load_data else data
+def plotData(data_file, use_pandas: bool, sequence: bool):
+ time_str = "$\mu$s"
+ show_grid = False
 
- if data.shape[1] < 2:
+ if use_pandas:
+ color = "red"
+ data = pd.read_csv(data_file)
+ # print(f"Pandas dataframe is:\n{data}")
+ # exit(1)
+ x = data.iloc[:, 1].to_numpy()
+ y = data.iloc[:, 3].to_numpy()
+ x_labels = (data.iloc[:, 1].to_numpy()).astype(int) # First column
+ # xy_pd = pd.DataFrame({"x" : x, "y": y})
+ # print(f"Pandas dataframe is:\n{data}\nx and y are:\n{xy_pd}")
+ # exit(1)
+ else:
+ color = "blue"
+ data = np.loadtxt(data_file, delimiter=',')
+ # data = np.genfromtxt(data_file, delimiter=',')
+ if data.shape[1] < 2:
+ print(f"[Error] (plotData): Data without two columns cannot be used for plotting x and y.")
+ exit(1)
+ x = data[:, 0] # First column
+ y = data[:, 1] * 1e6 # Second column
+ x_labels = data[:, 0].astype(int) # First column
+
+ y1 = y/x
+ y2 = y/(x*x)
+ y3 = y/(x*np.log(x))
+ if sequence:
+ x = np.arange(1, len(x_labels) + 1)
+
+ plt.figure(figsize=(10, 10)) # Optional: Adjusts the figure size
+ plt.subplot(131)
+ plt.plot(x, y1, color=color, marker='o', linestyle='-.') # Adjust markers and lines as needed
+ plt.title(f"Time ({time_str}) / N")
+ plt.xlabel("Cities") # Adjust with actual column name or description
+ # plt.ylabel(f"Time ({time_str})") # Adjust with actual column name or description
+ plt.grid(show_grid) # Optional: Adds a grid
+ if sequence:
+ axis = plt.gca()
+ plt.xticks(x, labels = x_labels, rotation='vertical')
+ plt.subplot(132)
+ plt.plot(x, y2, color=color, marker='o', linestyle='-.') # Adjust markers and lines as needed
+ plt.title(f"Time ({time_str}) / $N^2$")
+ plt.xlabel('Cities') # Adjust with actual column name or description
+ # plt.ylabel(f"Time ({time_str})") # Adjust with actual column name or description
+ plt.grid(show_grid) # Optional: Adds a grid
+ if sequence:
+ axis = plt.gca()
+ plt.xticks(x, labels = x_labels, rotation='vertical')
+ plt.subplot(133)
+ plt.plot(x, y3, color=color, marker='o', linestyle='-.') # Adjust markers and lines as needed
+ plt.title(f"Time ({time_str}) / $N \cdot log(N)$")
+ plt.xlabel('Cities') # Adjust with actual column name or description
+ # plt.ylabel(f"Time ({time_str})") # Adjust with actual column name or description
+ plt.grid(show_grid) # Optional: Adds a grid
+ if sequence:
+ axis = plt.gca()
+ plt.xticks(x, labels = x_labels, rotation='vertical')
+ if use_pandas:
+ plt.suptitle("Data from current implementation")
+ else:
+ plt.suptitle("Data from paper")
+ plt.show()
+
+def plotDataBoth(data_file1, data_file2, sequence: bool):
+ time_str = "$\mu$s"
+ show_grid = False
+
+ color1 = "blue"
+ data1 = np.loadtxt(data_file1, delimiter=',')
+ # data1 = np.genfromtxt(data_file1, delimiter=',')
+ if data1.shape[1] < 2:
  print(f"[Error] (plotData): Data without two columns cannot be used for plotting x and y.")
  exit(1)
- x = data[:, 0] # First column
- y = data[:, 1] # Second column
+ x1 = data1[:, 0] # First column
+ y1 = data1[:, 1] * 1e6 # Second column
+ x1_labels = data1[:, 0].astype(int) # First column
+
+ color2 = "red"
+ data2 = pd.read_csv(data_file2)
+ x2 = data2.iloc[:, 1].to_numpy()
+ y2 = data2.iloc[:, 3].to_numpy()
+ x2_labels = (data2.iloc[:, 1].to_numpy()).astype(int) # First column
 
- print(f"{x}\t{y}")
- plt.figure(figsize=(10, 6)) # Optional: Adjusts the figure size
- plt.plot(x, y, marker='o', linestyle='-') # Adjust markers and lines as needed
- plt.title('Number of Cites vs Time (seconds)')
- plt.xlabel('Cities') # Adjust with actual column name or description
- plt.ylabel('Time (s)') # Adjust with actual column name or description
  if sequence:
- x_labels = data[:, 0] # First column
+ x_labels = x1_labels
  x = np.arange(1, len(x_labels) + 1)
+ else:
+ x = x1
+
+ y1_1 = y1/x1
+ y1_2 = y1/(x1*x1)
+ y1_3 = y1/(x1*np.log(x1))
+
+ y2_1 = y2/x2
+ y2_2 = y2/(x2*x2)
+ y2_3 = y2/(x2*np.log(x2))
+
+ plt.figure(figsize=(10, 10)) # Optional: Adjusts the figure size
+ print(f"{x}\n{y1_1}")
+ print(f"{x}\n{y2_1}")
+ plt.subplot(131)
+ plt.plot(x, y1_1, color=color1, marker='o', linestyle='-.') # Adjust markers and lines as needed
+ plt.plot(x, y2_1, color=color2, marker='o', linestyle='-.') # Adjust markers and lines as needed
+ plt.title(f"Time ({time_str}) / N")
+ plt.xlabel("Cities") # Adjust with actual column name or description
+ # plt.ylabel(f"Time ({time_str})") # Adjust with actual column name or description
+ plt.grid(show_grid) # Optional: Adds a grid
+ if sequence:
+ axis = plt.gca()
+ plt.xticks(x, labels = x_labels, rotation='vertical')
+ print(f"{x}\n{y1_2}")
+ print(f"{x}\n{y2_2}")
+ plt.subplot(132)
+ plt.plot(x, y1_2, color=color1, marker='o', linestyle='-.') # Adjust markers and lines as needed
+ plt.plot(x, y2_2, color=color2, marker='o', linestyle='-.') # Adjust markers and lines as needed
+ plt.title(f"Time ({time_str}) / $N^2$")
+ plt.xlabel('Cities') # Adjust with actual column name or description
+ # plt.ylabel(f"Time ({time_str})") # Adjust with actual column name or description
+ plt.grid(show_grid) # Optional: Adds a grid
+ if sequence:
+ axis = plt.gca()
+ plt.xticks(x, labels = x_labels, rotation='vertical')
+ print(f"{x}\n{y1_3}")
+ print(f"{x}\n{y2_3}")
+ plt.subplot(133)
+ plt.plot(x, y1_3, color=color1, marker='o', linestyle='-.') # Adjust markers and lines as needed
+ plt.plot(x, y2_3, color=color2, marker='o', linestyle='-.') # Adjust markers and lines as needed
+ plt.title(f"Time ({time_str}) / $N \cdot log(N)$")
+ plt.xlabel('Cities') # Adjust with actual column name or description
+ # plt.ylabel(f"Time ({time_str})") # Adjust with actual column name or description
+ plt.grid(show_grid) # Optional: Adds a grid
+ if sequence:
  axis = plt.gca()
- plt.xticks(x, labels = x_labels)
- plt.grid(True) # Optional: Adds a grid
+ plt.xticks(x, labels = x_labels, rotation='vertical')
  plt.show()
 
 if __name__ == "__main__":
- plotData(DATA_FILE, load_data = True, sequence = False)
+ plotData(DATA_FILE1, False, sequence = True)
+ plotData(DATA_FILE2, True, sequence = True)
+ plotDataBoth(DATA_FILE1, DATA_FILE2, sequence = True)

utils.hpp

@@ -50,26 +50,20 @@ inline void printNewLine(int count = 1, std::ostream& stream = std::cout)
  stream << newline_str;
 }
 
-inline std::string subtituteStr(const std::string& text,
+inline std::string subtituteStr(std::string text,
  const std::string& pattern,
  const std::string& pattern_new)
 {
  typedef std::string::size_type ssize_type;
  const ssize_type n_old = pattern.length();
  const ssize_type n_new = pattern_new.length();
- ssize_type pos_begin = 0;
- ssize_type pos_found = text.find(pattern, pos_begin);
- if (pos_found == std::string::npos) {
- return text;
- }
- std::string text_new{ text };
- do {
- text_new = text_new.substr(0, pos_found) + pattern_new +
- text_new.substr(pos_found + n_old);
- pos_begin = pos_found + n_new;
- pos_found = text_new.find(pattern, pos_begin);
- } while (pos_found != std::string::npos);
- return text_new;
+ ssize_type pos_found = text.find(pattern, 0);
+ while (pos_found != std::string::npos) {
+ text = text.substr(0, pos_found) + pattern_new +
+ text.substr(pos_found + n_old);
+ pos_found = text.find(pattern, pos_found + n_new);
+ } ;
+ return text;
 }
 
 inline void printInfo(const std::string& msg, const std::string& fname = "")