Editorial: Puzzles are Still Challenging

This issue is dedicated to solving challenging puzzles. In the first article, Generating High-Score Crosswords Puzzles with Bi-Objective, Stochastic and Systematic Search, by Adi Botea and Vadim Bulitko, the Romanian Crosswords Competition Problem, also called Roco, is targeted. This puzzle is an NP-hard constraint optimization problem, where artificial intelligence (AI) used to lag substantially behind top human performance. The authors propose a search-based approach that can achieve a score of 190 for the competition year 2023, close to the top human score of 193. Subsequently, the article, Solving the Normal Magic Hexagon of Order Three: A “Smart Odometer” Treatment, by Kenneth Caviness, proposes an efficient algorithm to solve the normal magic hexagon with a side length of three.

In this issue, there is a report on the activities of the GameTable COST Action. On January 20 and 21, 2025, at Queen Mary University of London, a workshop took place focusing on game-playing AI, including search algorithms, knowledge representation, and reinforcement learning. Next, there is also a report on one of the AICUP competitions, the so-called Go Strength Imitation and Playstyle Recognition, which was held online from September 1, 2023, to January 2, 2024. The competition attracted an astonishing number of 508 teams.

Finally, there is a call for papers for a special issue on Adaptive and Intelligent Mobile Learning Systems for Personalized Learning and Gamification. The guest editors here will be Norshakirah Aziz, Hiroyuki Iida, and Abdullahi Abubakar Imam. They will look forward to your contributions.