No abstract

Sudoku is a complicated multidimensional mathematical structure with several applications in various computer science domains. 3D Sudoku, compared to 2D, has one more dimension that can potentially provide an extra edge to the different applications of this puzzle game. Various researchers have developed various types of 2D Sudoku solver using different methodologies. But there is very limited research in the area of developing 3D Sudoku solver. We have proposed two different solvers for solving 3D Sudoku puzzles of size 9×9×9. Both the solvers providing all possible solutions for a 3D Sudoku instance. 2D Sudoku puzzles are applied in different research domains with different purposes. Recently 3D structure of Sudoku is also applying in several areas to achieve more effectiveness compared to 2D Sudoku. As well as it can also be used to solve problem in 3D space. Again, solving an NP-complete puzzle by considering its 3D structure is a challenging job. Thus, we have endeavoured to achieve all possible solutions for a 3D Sudoku instance in this work. In the first version of our proposed algorithm all probable values for each blank cell have been computed and stored. Subsequently, few elimination-based methods have been used to reduce the number of probable values (if possible) for each blank cell. Finally, the solutions have been computed using the backtracking method. In the second version of our proposed algorithm, the nine 2D Sudoku puzzles lying in the xz-plane one above the other, which form the 3D puzzle, have been fed as the input. All possible solutions have been obtained for each of the nine puzzles. Then, the obtained solutions have been mapped to achieve one or more solutions for the 3D Sudoku instance. Thus, our proposed techniques provide a new approach for solving 3D Sudoku. In addition, applying the obtained solutions provides us with an advantage over 2D Sudoku, in solving problem of 3D space and where more data is required.

Noname manuscript No.

Due to their availability on commercial smartphones, WiFi, Bluetooth, and magnetometer are commonly utilized for indoor localization as indoor spaces are GPS deprived. Indoor localization falls into the category of data-intensive applications. In this domain, most of the recent solution approaches deploy MachineLearning (ML) and Deep Learning (DL) techniques on the data collected through the sensors. However, the publicly available benchmark datasets on indoor localization suffer from certain issues requiring complicated and customized data preprocessing techniques for each dataset for applying a common ML/DL technique. Thus, a fair comparison of the ML/DL methods for indoor localization datasets and hence, checking for the generality of a solution that spans across different indoor regions become infeasible. In this comparative study, we have investigated three key challenges of fingerprint datasets that should be addressed for real-life localization applications, namely, (i) repetitive site survey, (ii) device heterogeneity, (iii) granularity, and subregion-specific performance variation. To demonstrate how these attributes might impact localization performance, experimental analysis is done using five benchmark datasets. The novelty of the work is that it not only highlights the challenges but also analyse the feasibility of possible future directions to address these challenges through implementation results. Formulating the application of a generative adversarial network (GAN) to address the issue of repetitive site surveys has been discussed with implementation results.

No abstract