Scribd
Upload
146 views

Cable Trench Problem

The document describes the cable trench problem (CTP), which aims to minimize the cost of digging trenches and laying cable from a central hub to other nodes in a network. The CTP can be formulated as an integer program to find the optimal trenching and routing. While the problem of finding minimum spanning trees is in P, the CTP is NP-complete due to additional constraints. Previous work has developed heuristics and the problem can be solved using branch-cut-and-price algorithms implemented in frameworks like SYMPHONY. Future work includes generalizing the model and developing stronger cutting planes.

Uploaded by

Rurouni Kenshin
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
146 views

Cable Trench Problem

The document describes the cable trench problem (CTP), which aims to minimize the cost of digging trenches and laying cable from a central hub to other nodes in a network. The CTP can be formulated as an integer program to find the optimal trenching and routing. While the problem of finding minimum spanning trees is in P, the CTP is NP-complete due to additional constraints. Previous work has developed heuristics and the problem can be solved using branch-cut-and-price algorithms implemented in frameworks like SYMPHONY. Future work includes generalizing the model and developing stronger cutting planes.

Uploaded by

Rurouni Kenshin
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 12

Cable Trench Problem

Matthew V Galati Ted K Ralphs Joseph C Hartman


[email protected]

Department of Industrial and Systems Engineering Lehigh University, Bethlehem, PA

Cable Trench Problem p.1

Cable Trench Problem


The Cable Trench Problem (CTP) is that of minimizing the cost of digging trenches and laying cable for a communications network given a central hub. Let G = (N, A) be a connected digraph with specied depot 0 N . Dene cij as the cost/weight (typically distance) on arc (i, j ). Dene xed charge variables (trench) xij as to whether or not to dig a trench between nodes i and j . Dene ow variables (cable) yij as to the amount of cable to lay between nodes i and j .

Cable Trench Problem p.2

Node Routing
A node routing is a directed subgraph G of G satisfying the following properties:
G is (weakly) connected. The in-degree of each non-depot node is 1.

It is a spanning arborescence. There is a unique path from the depot to each demand point (vertex). Cost Measures (least cost routing)
Sum the lengths of all arcs in G . Sum the length of all paths from the depot. Some linear combination of these two.

Cable Trench Problem p.3

IP Formulation
M in
(i,j )A

cij (xij + xji ) + cij (yij + yji ) x( (N \ {i})) = 1 y ( (N \ {i})) y ( ({i})) = di i N \ {0} (1) i N \ {0} (2) (3) (4) (5)

s.t.

yij M xij (i, j ) A yij 0 (i, j ) A

xij {0, 1} (i, j ) A where:


(1) degree constraint (2) ow balance / demand constraint (3) capacity constraint

Cable Trench Problem p.4

Complexity
This node routing problem is NP-complete in general. Cable Trench Problem (, > 0)
= 0 Minimum Spanning Tree Problem. = 0 Shortest Paths Tree Problem.

Other special cases.


= 0 and x( ({i}) = 1 Traveling Salesman Problem (TSP). > 0 and x( ({i}) = 1 Variable Cost TSP (VCTSP). x( (V \ {0}) = x( ({0}) = k k -TSP.

Cable Trench Problem p.5

Sample Spanning Trees

1100 700

1100 700

1100 700

800 1400

800 1400

800 1400

400

2100

400

2100

400

2100

800 100 500 600 1200 600 100

800 100 500 600

800

500

1200

1200

1300 1300 22500 1300 22500

1300 1300 22500

1300

300 2100

900 2100

300

900 2100

300

900

1000 2500 1800 700

900 2500

1000

900 2500 1800

1000

900

1800 700

700

=0

/ = 10

/ = 0.1

/ = 0.001

Cable Trench Problem p.6

Complexity
Theorem 1 Among all minimum spanning trees nding the one that minimizes the path length between a particular set of vertices s and t is NP-Complete. Corollary 1 Among all minimum spanning trees nding the one that minimizes the total path length between a particular vertex s and all other vertices in V is NP-Complete. Corollary 2 The Cable Trench Problem is NP-Complete. Theorem 2 Among all shortest path trees rooted at s nding the one that minimizes the total edge length is in P.

Cable Trench Problem p.7

Previous Work
Vasko et. al - Kutztown University (to be published CAOR Nov 2001)
Heuristic upper bound for all values of / . The solution to CTP is a sequence of spanning trees such that as / increases, the total edge length strictly decreases each time another spanning tree becomes optimum and the total path length strictly increases. Total cost versus / is piecewise linear curve with strictly decreasing slopes.

Related Areas Fixed-Charge Network Flow Capacitated Network Design

Cable Trench Problem p.8

Valid Inequalities
Rounded Capacity Constraints xij d(S )/C
iS, j S

Flow Linking Constraints yij (C di )xij xij yij C di

yij y ( ({j })) xij dj Edge Cuts xij + xji 1


.5 0 1 .5 5 C=10 3

Cable Trench Problem p.9

Separation
Flow linking constraints and edge cuts can be included explicitly or separated in polynomial time. Separating rounded capacity constraints is NP-complete, but can be done effectively.

Cable Trench Problem p.10

Implementation
The implementation uses SYMPHONY, a parallel framework for branch, cut, and price (relative of COIN/BCP). In SYMPHONY, the user supplies:
the initial LP relaxation,separation subroutines, feasibility checker, and other optional subroutines.

SYMPHONY handles everything else. The source code and documentation are available from www.BranchAndCut.org Workshop on COIN/BCP (TB42) - Laszlo Ladanyi, Ted Ralphs
LP Relaxation SYMPHONY Internal Framework Separation routines Fesibility checker Optional routines

Solver

Cable Trench Problem p.11

Conclusions and Future Research


The ow linking constraints help to force integrality. The edge cuts also help impose structure and integrality. Future Research Generalizations of the model (different types of "trenches", different grades of "cable"). Better (more specic) cuts for the case where / is not extreme. Take advantage of connections to other models. Upper Bounds

Cable Trench Problem p.12

You might also like