A multiple coefficients trial method to solve combinatorial optimization problems for simulated-annealing-based ising machines

Kota Takehara, Daisuke Oku, Yoshiki Matsuda, Shu Tanaka, Nozomu Togawa

Research output: Chapter in Book/Report/Conference proceedingConference contribution

4 Citations (Scopus)

Abstract

When solving a combinatorial optimization problem with Ising machines, we have to formulate it onto an energy function of Ising model. Here, how to determine the penalty coefficients in the energy function is a great concern if it includes constraint terms. In this paper, we focus on a traveling salesman problem (TSP, in short), one of the combinatorial optimization problems with equality constraints. Firstly, we investigate the relationship between the penalty coefficient and the accuracy of solutions in a TSP. Based on it, we propose a method to obtain a TSP quasi-optimum solution, which is called multiple coefficients trial method. In our proposed method, we use an Ising machine to solve a TSP by changing a penalty coefficient every trial, the TSP solutions can converge very fast in total. Compared to naive methods using simulated-annealing-based Ising machines, we confirmed that our proposed method can reduce the total number of annealing iterations to 1/10 to 1/1000 to obtain a quasi-optimum solution in 32-city TSPs.

Original languageEnglish
Title of host publicationProceedings - 2019 IEEE 9th International Conference on Consumer Electronics, ICCE-Berlin 2019
EditorsGordan Velikic, Christian Gross
PublisherIEEE Computer Society
Pages64-69
Number of pages6
ISBN (Electronic)9781728127453
DOIs
Publication statusPublished - 2019 Sept
Event9th IEEE International Conference on Consumer Electronics, ICCE-Berlin 2019 - Berlin, Germany
Duration: 2019 Sept 82019 Sept 11

Publication series

NameIEEE International Conference on Consumer Electronics - Berlin, ICCE-Berlin
Volume2019-September
ISSN (Print)2166-6814
ISSN (Electronic)2166-6822

Conference

Conference9th IEEE International Conference on Consumer Electronics, ICCE-Berlin 2019
Country/TerritoryGermany
CityBerlin
Period19/9/819/9/11

Keywords

  • Combinatorial optimization problem
  • Ising machine
  • Penalty coefficient
  • Quadratic Unconstrained Binary Optimization
  • Traveling salesman problem

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Industrial and Manufacturing Engineering
  • Media Technology

Fingerprint

Dive into the research topics of 'A multiple coefficients trial method to solve combinatorial optimization problems for simulated-annealing-based ising machines'. Together they form a unique fingerprint.

Cite this