ABSTRACT
2048 is a simple and intriguing sliding block puzzle game that has been studied for several years. Many complex solvers, often developed using neural nets are available and capable of achieving very high scores. We are, however, interested in using only basic heuristics, the kind that could be conceivably employed by human players without the aid of computation. A common way to implement a 2048 solver involves searching the game tree for the best moves, choosing a move and scoring the game board using some evaluation functions. The choice in heuristic evaluation function can dramatically affect the moves chosen by the solver. Furthermore, two or more possible moves can frequently produce the same score as evaluated by the same heuristic function, requiring either a random choice, or the use of a secondary or back up evaluation function which itself in turn may produce a tie. In this paper, we test the effectiveness of several basic heuristics in a simple 2048 solver. In order to test these, we create a system that takes basic predefined heuristic evaluation functions as input parameters, generates compositions from these functions with certain rules, and automatically tests all of them with a specified number of games. We find that compositions of evaluation functions that maximize empty spaces and monotonicity of tiles on the board-especially those that prioritize high numbers of empty spaces above prioritizing higher monotonicity- perform the best out of all compositions that we test.
- Ahmed Abdelkader, Aditya Acharya, and Philip Dasler. 2016. 2048 Without New Tiles Is Still Hard. In FUN with Algorithms.Google Scholar
- Madhuparna Das and Goutam Paul. 2018. Analysis of the Game "2048" and its Generalization in Higher Dimensions. CoRR abs/1804.07393 (2018). arXiv:1804.07393 http://arxiv.org/abs/1804.07393Google Scholar
- David Eppstein. 2018. Making Change in 2048. CoRR abs/1804.07396 (2018). arXiv:1804.07396 http://arxiv.org/abs/1804.07396Google Scholar
- Wojciech Jaśkowski. 2016. Mastering 2048 with Delayed Temporal Coherence Learning, Multi-State Weight Promotion, Redundant Encoding and Carousel Shaping. CoRR abs/1604.05085 (2016). arXiv:1604.05085 http://arxiv.org/abs/1604.05085Google Scholar
- Stefan Langerman and Yushi Uno. 2018. Threes!, Fives, 1024!, and 2048 are hard. Theoretical Computer Science 748 (2018), 17 -- 27. FUN with Algorithms.Google ScholarCross Ref
- K. Matsuzaki. 2016. Systematic selection of N-tuple networks with consideration of interinfluence for game 2048. In 2016 Conference on Technologies and Applications of Artificial Intelligence (TAAI). 186--193.Google ScholarCross Ref
- Rahul Mehta. 2014. 2048 is (PSPACE) Hard, but Sometimes Easy. Electronic Colloquium on Computational Complexity (ECCC) 21 (2014), 116.Google Scholar
- Kazuto Oka and Kiminori Matsuzaki. 2016. Systematic Selection of N-Tuple Networks for 2048. In Computers and Games, Aske Plaat, Walter Kosters, and Jaap van den Herik (Eds.). Springer International Publishing, Cham, 81--92.Google Scholar
- P. Rodgers and J. Levine. 2014. An investigation into 2048 AI strategies. In 2014 IEEE Conference on Computational Intelligence and Games. 1--2.Google Scholar
- M. Szubert and W. Jaśkowski. 2014. Temporal difference learning of N-tuple networks for the game 2048. In 2014 IEEE Conference on Computational Intelligence and Games. 1--8.Google Scholar
- I-Chen Wu, Kun-Hao Yeh, Chao-Chin Liang, Chia-Chuan Chang, and Han Chiang. 2014. Multi-Stage Temporal Difference Learning for 2048. In Technologies and Applications of Artificial Intelligence, Shin-Ming Cheng and Min-Yuh Day (Eds.). Springer International Publishing, Cham, 366--378.Google Scholar
- Robert Xiao. 2014. What is the optimal algorithm for the game 2048? Retrieved September 14, 2018 from https://stackoverflow.com/a/22498940Google Scholar
- Robert Xiao, Wouter Vermaelen, and Petr MorÃąvek. 2018. 2048-AI. Retrieved September 14, 2018 from https://github.com/nneonneo/2048-aiGoogle Scholar
Index Terms
- Composition of basic heuristics for the game 2048
Recommendations
Expert review method in game evaluations: comparison of two playability heuristic sets
MindTrek '09: Proceedings of the 13th International MindTrek Conference: Everyday Life in the Ubiquitous EraThe expert review method is not yet widely adopted in game evaluations, although it has been used successfully in productivity software evaluations for years. In order to use the method effectively, there need to be playability heuristics that take into ...
Heuristics for Video Games Evaluation: How Players Rate Their Relevance for Different Game Genres According to Their Experience
IHM '13: Proceedings of the 25th Conference on l'Interaction Homme-MachineThis paper reports a study evaluating the relative relevance of evaluation heuristics for different game genres and players. 120 players (amateur or e-Sport players) were invited to assess the relevance of 47 heuristics for different game genre (Real ...
A comprehensive analysis of hyper-heuristics
Meta-heuristics such as simulated annealing, genetic algorithms and tabu search have been successfully applied to many difficult optimization problems for which no satisfactory problem specific solution exists. However, expertise is required to adopt a ...
Comments