Journal of Algorithms and Computation
Open Access

Implementation of Enhanced Multivariance Products Representation for Multiway Arrays

Document Type : Research Paper

Author

Cosar Gozukirmizi

Istanbul Beykent University

https://doi.org/10.22059/jac.2025.394087.1228
Abstract
Multiway arrays (also called multilinear arrays, tensors, folded arrays) can be decomposed by Enhanced Multivariance Products Representation (EMPR). In this work, Enhanced Multivariance Products Representation for multiway arrays is implemented in Julia programming language. A program for computing all the terms of the EMPR for multiway arrays is written. We have utilized the ITensor library which makes the multiway array operations easier to understand and implement. The function in the program is able to decompose any multiway array of floating point numbers. The number of ways of the original array can be any positive integer. However, large multiway arrays would cause problems related to the fact that the multiway arrays may not fit on computer's memory.

Keywords

high dimensional model representation
multidimensional array decomposition
tensor network diagram
Julia programming
ITensor library
Journal of Algorithms and Computation
Volume 57, Issue 1
August 2025
Pages 96-111

Home

Guide for Authors

Submit Manuscript

Reviewers

Contact Us