Publications by Martin Butler

Table Talk Tuning is a table-assignment problem drawn from a real wedding seating case. Alongside capacities, grouping and separation requests, table slack, and balance constraints, the model asks that every guest have at least one topic with enough support at their table to sustain a conversation.

The paper presents a MiniZinc model, three synthetic instance families from 10 to 150 guests, a reproducible benchmark pipeline, and a native Gecode model for comparing objective-aware search. The experiments show that the conversation terms still change the returned seatings in the paired ablation cases. The native experiments show that scalar, lexicographic, and sequential handling of the same objective hierarchy changes search behaviour.

Kingdomino is introduced as an interesting game for studying game playing: the game is multiplayer (4 independent players per game); it has a limited game depth (13 moves per player); and it has limited but not insignificant interaction among players.

Several strategies based on locally greedy players, Monte Carlo Evaluation (MCE), and Monte Carlo Tree Search (MCTS) are presented with variants. We examine a variation of UCT called progressive win bias and a playout policy (Player-greedy) focused on selecting good moves for the player. A thorough evaluation is done showing how the strategies perform and how to choose parameters given specific time constraints. The evaluation shows that surprisingly MCE is stronger than MCTS for a game like Kingdomino.

All experiments use a cloud-native design, with a game server in a Docker container, and agents communicating using a REST-style JSON protocol. This enables a multi-language approach to separating the game state, the strategy implementations, and the coordination layer.

Sheet metal cutting using lasers is ubiquitous in the industry, and is used to produce everything from home decorations to excavator scoops. Metal waste is costly for the industry, both in terms of money, but also in terms of an increased environmental footprint. Tomologic develops a unique optimisation system that can reduce this waste drastically. This paper presents a CP approach to the Laser Cutting Path Planning Problem (LCPPP), a very hard important sub problem within the Tomologic optimisation system. A solution to the LCPPP is, given a packing of some details on a metal sheet, an ordering of the cuts necessary to separate the details from the sheet. The problem is complicated by physical factors such as heat from the laser beam, or details moving or flexing. In the paper, we explain the problem in detail and present our CP approach that we developed for solving the problem. The possibility (in CP) of custom search heuristics turned out to be crucial to be able to solve the problem efficiently, as these could be made to guide the search to good first solutions.