### Abstract

We consider the problem of balancing the load among several service-providing facilities, while keeping the total cost low. Let D be the underlying demand region, and let p
_{1},p
_{m} be m points representing m facilities. We consider the following problem: Subdivide D into m equal-area regions R
_{1},R
_{m} , so that region R
_{i} is served by facility p
_{i} , and the average distance between a point q in D and the facility that serves q is minimal. We present constant-factor approximation algorithms for this problem, with the additional requirement that the resulting regions must be convex. As an intermediate result we show how to partition a convex polygon into m equal-area convex subregions so that the fatness of the resulting regions is within a constant factor of the fatness of the original polygon. In fact, we prove that our partition is, up to a constant factor, the best one can get if one's goal is to maximize the fatness of the least fat subregion. We also discuss the structure of the optimal partition for the aforementioned load balancing problem: indeed, we argue that it is always induced by an additive-weighted Voronoi diagram for an appropriate choice of weights.

Original language | English (US) |
---|---|

Pages (from-to) | 318-336 |

Number of pages | 19 |

Journal | Algorithmica (New York) |

Volume | 54 |

Issue number | 3 |

DOIs | |

State | Published - Jul 2009 |

### Fingerprint

### Keywords

- Additive-weighted Voronoi diagram
- Approximation algorithms
- Fat partitions
- Fatness
- Geometric optimization
- Load balancing

### ASJC Scopus subject areas

- Computer Science(all)
- Computer Science Applications
- Applied Mathematics

### Cite this

*Algorithmica (New York)*,

*54*(3), 318-336. https://doi.org/10.1007/s00453-007-9125-3

**Minimum-Cost Load-Balancing Partitions.** / Aronov, Boris; Carmi, Paz; Katz, Matthew J.

Research output: Contribution to journal › Article

*Algorithmica (New York)*, vol. 54, no. 3, pp. 318-336. https://doi.org/10.1007/s00453-007-9125-3

}

TY - JOUR

T1 - Minimum-Cost Load-Balancing Partitions

AU - Aronov, Boris

AU - Carmi, Paz

AU - Katz, Matthew J.

PY - 2009/7

Y1 - 2009/7

N2 - We consider the problem of balancing the load among several service-providing facilities, while keeping the total cost low. Let D be the underlying demand region, and let p 1,p m be m points representing m facilities. We consider the following problem: Subdivide D into m equal-area regions R 1,R m , so that region R i is served by facility p i , and the average distance between a point q in D and the facility that serves q is minimal. We present constant-factor approximation algorithms for this problem, with the additional requirement that the resulting regions must be convex. As an intermediate result we show how to partition a convex polygon into m equal-area convex subregions so that the fatness of the resulting regions is within a constant factor of the fatness of the original polygon. In fact, we prove that our partition is, up to a constant factor, the best one can get if one's goal is to maximize the fatness of the least fat subregion. We also discuss the structure of the optimal partition for the aforementioned load balancing problem: indeed, we argue that it is always induced by an additive-weighted Voronoi diagram for an appropriate choice of weights.

AB - We consider the problem of balancing the load among several service-providing facilities, while keeping the total cost low. Let D be the underlying demand region, and let p 1,p m be m points representing m facilities. We consider the following problem: Subdivide D into m equal-area regions R 1,R m , so that region R i is served by facility p i , and the average distance between a point q in D and the facility that serves q is minimal. We present constant-factor approximation algorithms for this problem, with the additional requirement that the resulting regions must be convex. As an intermediate result we show how to partition a convex polygon into m equal-area convex subregions so that the fatness of the resulting regions is within a constant factor of the fatness of the original polygon. In fact, we prove that our partition is, up to a constant factor, the best one can get if one's goal is to maximize the fatness of the least fat subregion. We also discuss the structure of the optimal partition for the aforementioned load balancing problem: indeed, we argue that it is always induced by an additive-weighted Voronoi diagram for an appropriate choice of weights.

KW - Additive-weighted Voronoi diagram

KW - Approximation algorithms

KW - Fat partitions

KW - Fatness

KW - Geometric optimization

KW - Load balancing

UR - http://www.scopus.com/inward/record.url?scp=67349206754&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=67349206754&partnerID=8YFLogxK

U2 - 10.1007/s00453-007-9125-3

DO - 10.1007/s00453-007-9125-3

M3 - Article

VL - 54

SP - 318

EP - 336

JO - Algorithmica

JF - Algorithmica

SN - 0178-4617

IS - 3

ER -