Potential game distributed machine learning solution method of large-scale sudoku problem

Abstract

The invention discloses a game theory optimization method for carrying out distributed solution on a sudoku problem and gives physical game realization of the sudoku problem. The method comprises the following steps: (1), building a utility function and proving that the sudoku problem can be converted into a potential game model; (2) optimizing the states of participants step by step by using learning motivation so that the states of the participants reach the optimal states, namely Nash equilibrium point of the potential game.

Description

technical field

[0001] The invention adopts a potential game distributed machine learning solution method of a large-scale Sudoku problem, and provides a physical game realization of the Sudoku problem, belonging to the field of multi-agent intelligent cooperation. Background technique

[0002] sudoku problem

[0003] Sudoku has been described as the Rubik's Cube of the twenty-first century. Sudoku is a popular and seemingly addictive puzzle that has been played in many parts of the world. The goal of Sudoku is simple: put The square is divided into n n different grids, the purpose is to fill each square so that the following three conditions are met:

[0004] (1) The numbers filled in the squares of each row are from 1 to n 2 can only appear once

[0005] (2) The numbers filled in the squares of each column are from 1 to n 2 can only appear once

[0006] (3) The numbers filled in the squares in each grid range from 1 to n 2 can only appear once

[0007] The Su...

Images