Heuristic solver for tsp

Author: philipandreadis

Algorithm_tests/graphtheory_tests/test_NN.py

@@ -0,0 +1,69 @@
+import unittest
+import sys
+
+# Import from different folder
+sys.path.append("Algorithms/graphtheory/nearest-neighbor-tsp/")
+
+import NearestNeighborTSP
+
+
+class TestNN(unittest.TestCase):
+    def setUp(self):
+        self.G1 = [[0,3,-1],[3,0,1],[-1,1,0]]
+        self.correct_path1 = [0,1,2,0]
+
+        # No possible solution for this one so its a dead end
+        self.G2 = [[0, 2, -1,-1,-1], [2, 0,5,1,-1], [-1, 5, 0, -1, -1],[-1, 1, -1, 0, 3], [-1, -1, -1, 3, 0]]
+        self.correct_path2 = [0,1,3,4]
+
+        # No possible solution for this one so its a dead end
+        self.G3 = [[0, 2, -1,-1,-1], [2, 0,5,1,-1], [-1, 5, 0, -1, -1],[-1, 1, -1, 0, -1], [-1, -1, -1, -1, 0]]
+        self.correct_path3 = [0, 1, 3]
+
+        # Multiple possible solutions
+        self.G4 = [[0,1,1,1],[1,0,1,1],[1,1,0,1],[1,1,1,0]]
+        self.correct_path4 = [0, 1, 2, 3, 0]
+
+
+    # adjacency matrix of a graph for testing
+    adjMatrix = [[0,2,5,-1,3],[2,0,2,4,-1],[5,2,0,5,5],[-1,4,5,0,2],[3,-1,5,2,0]]
+    # correct rank of each node's neighbors
+    correctNeighbors = [[1,4,2],[0,2,3],[1,0,3,4],[4,1,2],[3,0,2]]
+
+
+    def test_0_rankNeighbors(self):
+        for i in range(0,4):
+            self.assertEqual(NearestNeighborTSP.rankNeighbors(i, self.adjMatrix), self.correctNeighbors[i], "Check if order is different.")
+
+
+    def test_1_nnTSP(self):
+        path=NearestNeighborTSP.nnTSP(self.adjMatrix)
+        # Test if path is null
+        self.assertIsNotNone(path,"Output is empty")
+        # Test if path is not complete
+        self.assertEqual(len(path),len(self.adjMatrix)+1,"Path in incomplete")
+
+
+    def test_linear_graph(self):
+        #print(NearestNeighbor.nnTSP(self.G2))
+        path = NearestNeighborTSP.nnTSP(self.G1)
+        self.assertEqual(path,self.correct_path1)
+
+
+    def test_simple_graph(self):
+        path = NearestNeighborTSP.nnTSP(self.G2)
+        self.assertEqual(path,self.correct_path2)
+
+
+    def test_disconnected_graph(self):
+        path = NearestNeighborTSP.nnTSP(self.G3)
+        self.assertEqual(path, self.correct_path3)
+
+
+    def test_complete_graph(self):
+        path = NearestNeighborTSP.nnTSP(self.G4)
+        self.assertEqual(path, self.correct_path4)
+
+if __name__ == '__main__':
+    print("Running Nearest Neighbor TSP solver tests:")
+    unittest.main()

Algorithms/graphtheory/nearest-neighbor-tsp/NearestNeighborTSP.py

@@ -0,0 +1,126 @@
+"""
+Author: Philip Andreadis
+e-mail: [email protected]
+
+    This script implements a simple heuristic solver for the Traveling Salesman Problem.
+    It is not guaranteed that an optimal solution will be found.
+
+    Format of input text file must be as follows:
+    1st line - number of nodes
+    each next line is an edge and its respective weight
+
+    The program is executed via command line with the graph in the txt format as input.
+
+"""
+
+import time
+import sys
+
+
+
+"""
+This function reads the text file and returns a 2d list which represents 
+the adjacency matrix of the given graph
+"""
+def parseGraph(path):
+  # Read number of vertices and create adjacency matrix
+  f = open(path, "r")
+  n = int(f.readline())
+  adjMatrix = [[-1 for i in range(n)] for j in range(n)]
+  #Fill adjacency matrix with the correct edges
+  for line in f:
+    edge = line.split(" ")
+    edge = list(map(int, edge))
+    adjMatrix[edge[0]][edge[1]] = edge[2]
+    adjMatrix[edge[1]][edge[0]] = edge[2]
+  for i in range(len(adjMatrix)):
+    adjMatrix[i][i] = 0
+  return adjMatrix
+
+
+
+"""
+Returns all the neighboring nodes of a node sorted based on the distance between them.
+"""
+def rankNeighbors(node,adj):
+    sortednn = {}
+    nList = []
+    for i in range(len(adj[node])):
+        if adj[node][i]>0:
+            sortednn[i] = adj[node][i]
+    sortednn = {k: v for k, v in sorted(sortednn.items(), key=lambda item: item[1])}
+    nList = list(sortednn.keys())
+    return nList
+
+
+"""
+Function implementing the logic of nearest neighbor TSP.
+Generate two lists a and b, placing the starting node in list a and the rest in list b.
+While b is not empty append to a the closest neighboring node of the last node in a and remove it from b.
+Repeat until a full path has been added to a and b is empty.
+Returns list a representing the shortest path of the graph.
+"""
+def nnTSP(adj):
+    nodes = list(range(0, len(adj)))
+    #print(nodes)
+    weight = 0
+    global length
+    # Starting node is 0
+    a = []
+    a.append(nodes[0])
+    b = nodes[1:]
+    while b:
+        # Take last placed node in a
+        last = a[-1]
+        # Find its nearest neighbor
+        sortedNeighbors = rankNeighbors(last,adj)
+        # If node being checked has no valid neighbors and the path is not complete a dead end is reached
+        if (not sortedNeighbors) and len(a)<len(nodes):
+            print("Dead end!")
+            return a
+        flag = True
+        # Find the neighbor that has not been visited
+        for n in sortedNeighbors:
+            if n not in a:
+                nextNode = n
+                flag = False
+                break
+        # If all neighbors of node have been already visited a dead end has been reached
+        if flag:
+            print("Dead end!")
+            return a
+        a.append(nextNode)
+        b.remove(nextNode)
+        # Add the weight of the edge to the total length of the path
+        weight = weight + adj[last][nextNode]
+    # Finishing node must be same as the starting node
+    weight = weight + adj[a[0]][a[-1]]
+    length = weight
+    a.append(a[0])
+    return a
+
+
+if __name__ == "__main__":
+  # Import graph text file
+  filename = sys.argv[1]
+  print("~Nearest Neighbor~")
+  start = time.time()
+  graph = parseGraph(filename)
+  n = len(graph)
+  length = 0
+
+  # Print adjacency matrix
+  print("Adjacency matrix of input graph:")
+  for i in range(len(graph)):
+      print(graph[i])
+
+  start = time.time()
+
+  path = nnTSP(graph)
+
+  finish = time.time()-start
+
+  # Output
+  print("Path:",path)
+  print("Length",length)
+  print("Time:",finish)

Algorithms/graphtheory/nearest-neighbor-tsp/graph.txt

@@ -0,0 +1,9 @@
+5
+0 1 2
+0 2 5
+0 4 3
+1 2 2
+1 3 4
+2 3 5
+2 4 5
+3 4 2