30.01.06 Michael Heusch - IntCP 2006 Modeling and solving of a radio antennas deployment support application with discrete and interval constraints.

Michael Heusch - IntCP 2006

30.0

1.0 6

Modeling and solving of a radio antennasdeployment support application with discreteand interval constraints


30.0

1.0 6

Outline of the talk

Presentation of the application

Modeling with discrete and interval constraints

Defining search heuristics

Modeling the problem with the distn constraint

Experimental results on solving the progressive deployment problem


30.0

1.0 6

Presentation of the LocRLFAP

Informal description of the de radio antennas deployment problem :

Constraints involved : Distance between

frequencies depends on distance between antennas


30.0

1.0 6

Minimal and maximal distances between antennas

Presentation of the LocRLFAP

Informal description of the de radio antennas deployment problem :

Difficulties : Hybrid combinatorial

optimisation problem

non-linear continuous constraints

Constraints involved : Distance between

frequencies depends on distance between antennas


30.0

1.0 6

Specification of the problem

Formulation as a constrained optimisation problem: Data

Fixed set of antennas (transmitter-receiver)

Dispatched on n sites {P1, … , Pn}

The links to establish is known in advance

Variables of the problem: A solution associates one frequency to each antenna and a position to

each site

Pi = (Xi,Yi): Position of a site

fi,j : frequency allocated to the link from Pi to Pj

Optimisation problem: Minimise the maximal frequency used


30.0

1.0 6

Constraints of the problem

Constraints of the problem

discrete constraints:

Compatibility between antennas

Forbidden frequencies

continuous constraints

Maximum distance between antennas (range)

Minimum distance between the antennas (security, interference)

mixed constraints

Compatibility between the allocation and the deployment


30.0

1.0 6

Comparing the RLFAP/LocRLFAP with 5 sites

RLFAP LocRLFAP


30.0

1.0 6

Comparing the RLFAP/LocRLFAP with 5 sites

RLFAP LocRLFAP

dist² (Si,Sj) = Σi (Xi - Xj)²


30.0

1.0 6

Comparing the RLFAP/LocRLFAP


30.0

1.0 6

Hybrid solving with collaborating solvers

Original approach

Modeling with the finite domain constraint solver Eclair

Full discretization of the problem

Modeling three types of constraints

Discrete constraints

Continuous constraints

Mixed constraints


30.0

1.0 6

Discrete constraints

Co-site transmitter-receiver interference constraints:

Duplex distance constraints for each bidirectional link

Forbidden portions in the frequency range


30.0

1.0 6

Continuous and mixed constraints

Elementary continuous constraints:

dist²(Pi,Pj) > mij² , for all i<j

dist²(Pi,Pj) < Mij² , if there exists a radio link between Pi and Pj

Mixed constraints: Compatibility constraints

If dist(Pi,Pj)< d1, great interference

If d1 <= dist(Pi,Pj)< d2, limited interference

Expression with elementary constraints { dist(Pi,Pj)< d1 } v { |fik-fjl| > Δ1 }, (i,j,k), i≠j, i≠k, j≠k

{ dist(Pi,Pj)< d2 } v { |fik-fjl| > Δ2 }, (i,j,k), i≠j, i≠k, j≠k

d1

d2


30.0

1.0 6

Test set

Full deployment of networks with 5 to 10 sites

RLFAP

LocRLFAP


30.0

1.0 6

Progressive deployment of networks with 9 and 10 sites

P

P

P

P

P

P

P

P P

P


30.0

1.0 6

Solving with elementary constraints

Full deployment in both models


30.0

1.0 6

Improvements to the search algorithm

Usage of a naïve Branch & Bound with:

Distinction of the type of variables

The problem is under-constrained on positions

Branch on disjunctions?

Branch first on constraints entailing a strong interdistance?

Variable selection heuristics

minDomain

min(dom/deg)

minDomain+maxConstraints


30.0

1.0 6

Results with minDomain+maxConstraints

9 sites 10 sites

99% of the backtracks are performed on the continuous part of the search tree

A bit less backtracks on the hybrid model

Hybrid solving is 1 to 3 times slower

Progressive deployment in both models


30.0

1.0 6

Introducing the distn global constraint

distn ([P1, … , Pn], V)

Pi = Xi x Yi : Cartesian coordinates of the point pi

V i,j : distance to maintain between Pi and Pj

distn(p1, … , pn], v)

satisfied if and only if

dist(pi,pj) = vi,j

Filtering algorithm uses geometric approximation techniques


30.0

1.0 6

Applications of the constraint

Molecular conformation

Robotics

Antennas deployment


30.0

1.0 6

Using distn in the model

Second formulation of the problem with the global constraint:

Simple continuous constraints

Introduction of a matrix {Vi,j} of distance variables:

Domain(Vi,j)=[mi,j , Mi,j]

Expression of the set of min and max distance constraints:distn([P1, … , Pn], V)

Expression of the mixed « distant compatibility » disjunctions distn([P1, … , Pn], V)

{ Vij<d 1 } v { |fik-fjl| > Δ 1 }, (i,j,k), i≠j, i≠k, j≠k

{ Vij<d 2 } v { |fik-fjl| > Δ 2 }, (i,j,k), i≠j, i≠k, j≠k


30.0

1.0 6

Results using distn (9 sites)

Simple heuristics Advanced heuristics

hybrid model / discrete model comparison:

1.8 times slower

1.5 times more backtracks

Similar performance of both models

wrt. simple model, distn divides by 2 the nb. of backtracks


30.0

1.0 6

Results using distn (10 sites)

Simple heuristics Advanced heuristics

hybrid model / discrete model comparison:

4 additional instances are solved

Performance on the solved instances:

• 63% less backtracks

All instances are solved


30.0

1.0 6

Quality of solutions

9 sites 10 sites


30.0

1.0 6

Conclusion and perspectives

We showed the relevance of coupling discrete and continuous constraints Obtain solution of greater quality

Better performance when solving

Independence w.r.t. the discretization step

Validation on one industrial application

Key points Definition of appropriate search heuristics

Usage of the distn global constraint


30.0

1.0 6

Perspectives on the application

Validation on instances of greater size

Take forbidden zone constraints into account

Provide deployment zones using polygons


30.0

1.0 6

Other approaches for solving the RLFAP

Other approaches for solving the classical RLFAP

Graph coloring

Branch & Cut

CP LDS [Walser – CP96] Russian Doll Search [Schiex et. al - CP97]

Heuristics Tabou [Vasquez – ROADEF 2001] Simulated annealing, evolutionary algorithms…

Motivations for an approach using CP

Robustness wrt modification of the constraints of the problem


30.0

1.0 6

Sketch of distn’s filtering algorithm


30.0

1.0 6

Filtering algorithm on polygons

Method using polygons for representing domains

Theorem by K. Nurmela et P. Östergård (1999)

M. Markót et T. Csendes: A New Verified Optimization Technique for the ``Packing Circles in a Unit Square'' Problems. SIAM Journal of Optimization, 2005

pi2

pik-1

pik

pi1

Pi

Pj


30.0

1.0 6


P1

P2


30.0

1.0 6


-

-

+

+

+

+

P1

P2


30.0

1.0 6


P1

P2


30.0

1.0 6

Interval extension of the algorithm

P1

P2


30.0

1.0 6

Filtering algorithm of distn

P1

P2

Adjusting bounds of the distance variables

30.01.06 Michael Heusch - IntCP 2006 Modeling and solving of a radio antennas deployment support application with discrete and interval constraints.

Documents

michael heusch intcp

problem discrete constraints

interval constraints

duplex distance constraints

deployment slide

problem modeling

antennas presentation

sites rlfaplocrlfap