METROPOLIZED FOREST RECOMBINATION FOR MONTE CARLO SAMPLING OF GRAPH PARTITIONS
| dc.contributor.author | Autry, E | |
| dc.contributor.author | Carter, D | |
| dc.contributor.author | Herschlag, GJ | |
| dc.contributor.author | Hunter, Z | |
| dc.contributor.author | Mattingly, JC | |
| dc.date.accessioned | 2024-02-05T15:13:08Z | |
| dc.date.available | 2024-02-05T15:13:08Z | |
| dc.date.issued | 2023-08-01 | |
| dc.description.abstract | We develop a new Markov chain on graph partitions that makes relatively global moves yet is computationally feasible to be used as the proposal in the Metropolis-Hastings method. Our resulting algorithm is able to sample from a specified measure on partitions or spanning forests. Being able to sample from a specified measure is a requirement of what we consider as the gold standard in quantifying the extent to which a particular map is a gerrymander. Our proposal chain modifies the recently developed method called recombination (ReCom), which draws spanning trees on joined partitions and then randomly cuts them to repartition. We improve the computational efficiency by augmenting the statespace from partitions to spanning forests. The extra information accelerates the computation of the forward and backward proposal probabilities which are required for the Metropolis-Hastings algorithm. We demonstrate this method by sampling redistricting plans on several measures of interest and find promising convergence results on several key observables of interest. We also explore some limitations in the measures that are efficient to sample from and investigate the feasibility of using parallel tempering to extend this space of measures. | |
| dc.identifier.issn | 0036-1399 | |
| dc.identifier.issn | 1095-712X | |
| dc.identifier.uri | ||
| dc.language | en | |
| dc.publisher | Society for Industrial & Applied Mathematics (SIAM) | |
| dc.relation.ispartof | SIAM Journal on Applied Mathematics | |
| dc.relation.isversionof | 10.1137/21M1418010 | |
| dc.rights.uri | ||
| dc.subject | Markov chain Monte Carlo | |
| dc.subject | balanced graph partitioning | |
| dc.subject | Metropolis-Hastings | |
| dc.subject | spanning trees | |
| dc.title | METROPOLIZED FOREST RECOMBINATION FOR MONTE CARLO SAMPLING OF GRAPH PARTITIONS | |
| dc.type | Journal article | |
| duke.contributor.orcid | Herschlag, GJ|0000-0001-5443-6449 | |
| duke.contributor.orcid | Mattingly, JC|0000-0002-1819-729X | |
| pubs.begin-page | 1366 | |
| pubs.end-page | 1391 | |
| pubs.issue | 4 | |
| pubs.organisational-group | Duke | |
| pubs.organisational-group | Trinity College of Arts & Sciences | |
| pubs.organisational-group | Mathematics | |
| pubs.organisational-group | Statistical Science | |
| pubs.publication-status | Published | |
| pubs.volume | 83 |
Files
Original bundle
- Name:
- 21m1418010.pdf
- Size:
- 2.74 MB
- Format:
- Adobe Portable Document Format
- Description:
- Published version