# Matrix model
A matrix model is a model in which a matrix or multiple matrices are the dynamical objects.
## Refs
- 't Hooft expansion
- [[1974#'t Hooft]]
- topological recursion
- [[1983#Migdal]]: so-called loop equations determine all order results in the large-$N$ expansion of the correlators, where $N$ is the size of the matrix
- [[2004#Eynard]] and [[2007#Eynard, Orantin]]: topological recursion relations
- topological matrix models (no double scaling)
- [[1992#Kontsevich]]
- doubled scaled matrix integral
- [[1990#Brezin, Kazakov]]
- [[1990#Douglas, Shenker]]
- [[1990#Gross, Migdal]]
- $c=1$ string
- [[2003#McGreevy, Verlinde]]: matrix model for $c=1$ strings as the open string field theory for an infinite number of D0-branes
- [[2003#Klebanov, Maldacena, Seiberg]]: the double-scaled matrix model can be understood as arising from [[0652 ZZ brane|ZZ branes]]
- Hilbert space in ETH matrix model
- [[2025#Miyaji, Mori, Okuyama]]
- misc topics
- [[Raman2021]][](https://arxiv.org/pdf/2110.06643.pdf): replica
- [[KarLamprouMarteauRosso2022]][](https://arxiv.org/pdf/2208.05974.pdf): flat space dual
## The Vandermonde determinant
A matrix integral for a Hermitian matrix $H$ given by$\mathcal{M}_N=\int_{\mathbb{R}^{N^2}}[\mathrm{d} H]\, \mathrm{e}^{-N \operatorname{tr} V(H)}$can be recast by diagonalising the matrix into$\mathcal{M}_N=\int_{\mathbb{R}^N} \prod_{i=1}^N \mathrm{~d} \lambda_i \mathrm{e}^{-N^2 S[\lambda]},$where$S[\lambda]=\frac{1}{N} \sum_{i=1}^N V\left(\lambda_i\right)-\frac{1}{N^2} \sum_{i \neq j} \log \left|\lambda_i-\lambda_j\right|.$The second term is known as the Vandermonde, which comes from the Jacobian of the diagonalisation.
## Double scaling limit
The double scaling limit involves taking the size of the matrix to infinity. At the same time, one focused on the edge of the spectrum.
## Related topics
- [[0471 String-matrix duality]]
- [[0579 Random matrix theory]]