Complete solution of eight puzzle problem using BFS in CUDA environment

Abstract

The eight puzzle problem is the largest completely solvable problem of n∗n sliding puzzle problems. It is combinatorial in nature, but there is a large problem space of 9! /2. Objective of this work is to find the complete solution of eight puzzle problem i.e. examining all the permutations for solvability. Using Breadth First search (BFS) graph traversal we can reach the solution for a definite goal. The parallel algorithm is capable of providing us with much more faster solution using Compute Unified Device Architecture (CUDA). CUDA facility enables us to use best possible available computation power of GPU (Graphics Processing Unit). In this paper, we present fast implementation of common graph operation like breadth-first search to find out complete solution of eight puzzle problem on the GPU using the CUDA programming model. Our implementations exhibit better performance. The availability and spread of GPUs to desktops and laptops make them ideal candidates to accelerate graph operations over the CPU-only implementations.