The document discusses Steiner trees, which are trees that connect a set of points while allowing additional intermediate points called Steiner points. Some key points:
1) Adding Steiner points can significantly reduce the total length needed to connect points compared to just connecting the points directly. For example, adding one Steiner point to connect three points in a triangle can reduce the length by over 13%.
2) Steiner trees have applications in areas like circuit board design where they help minimize the total wiring or pathway length needed.
3) Finding optimal Steiner trees is computationally difficult, so heuristics like depth-first search are often used to find approximate solutions, especially for larger point sets.
2. The Steiner problem dates back to Pierre Fermat. Even Napolean has worked on this problem. Consider three points V 1 , V 2 , and V 3 . The shortest network interconnecting these three points is the sum of the two shortest sides of the triangle. If one includes a fourth point S 1 in the plane of the triangle, one can find a much shorter interconnecting network. V 1 V 2 V 3 V 1 V 2 V 3 S 1
3. If the three points in question are vertices of an equilateral triangle with side of length 2 kms, one can make some simple calculations such as: Length = 4 kms 2/3 (  3) Length = 3(2/3)  3 = 3.4641 kms a saving of about 13.4%  3 V 1 V 2 V 3 V 1 V 2 V 3 S 1
4. If it costs about Rs. 3 lakhs to lay one km of road this is a saving of 0.134 * 300,000 * 4 = Rs. 160,800 on an initial total cost of Rs. 12 lakhs.
5. Steiner points have some properties Exactly three edges meet at every Steiner vertex. The angles between the edges meeting at a Steiner vertex is 120 degrees. If there are N vertices then we can use a maximum of N-2 Steiner vertices.
6. MST length = 1 + ï€ ïƒ– 3 SMT length is 2.6458 A saving of 3.1588% 1  3 2 (-0.2857, 0.2474)
7. Some triangles are such that a Steiner point does not help 2 2 MST length = SMT length = 4 No savings!
27. Depth First Search of point set – The first Steiner point Heuristic for large point set
28. Depth First Search of point set – The second Steiner point Heuristic for large point set
29. Depth First Search of point set – The third Steiner point Heuristic for large point set
30. Depth First Search of point set – The fourth Steiner point Heuristic for large point set
31. Depth First Search of point set – Return to root node Heuristic for large point set
32. Depth First Search of point set – First Steiner vertex Heuristic for large point set
33. Depth First Search of point set – Second Steiner vertex Heuristic for large point set
34. Depth First Search of point set – Third Steiner vertex Heuristic for large point set
35. Depth First Search of point set – Fourth Steiner vertex Heuristic for large point set
36. Depth First Search of point set – Fifth Steiner vertex Heuristic for large point set
37. Steiner Tree and Minimum Spanning Tree for a small random point set in unit cube Minimal Interconnecting Tree for data without any structure
38. Steiner Tree and Minimum Spanning Tree for a medium sized random point set in unit cube Minimal Interconnecting Tree for data without any structure
39. Steiner Tree and Minimum Spanning Tree for Hundred random points in unit cube Minimal Interconnecting Tree for data without any structure
40. Steiner Tree and Minimum Spanning Tree for 500 random points in unit cube Minimal Interconnecting Tree for data without any structure