![]() For example, Player 1 can know that Player 2 knows that the Red Diamond Solitaire is in the game: K 1(K 2(RDS)). The agents are also able to use higher-level knowledge, which is knowledge about knowledge. Due to Axiom A3 of the S5 (m) logic system, this implies that this Gem-Card is in the game: K 1(RDS) → RDS. For example, if Player 1 knows that the Red Diamond Solitaire is in the game, this is represented as K 1(RDS). First-level knowledge is a combination of a player that the knowledge belongs to and the Gem-Card that this player knows is in the game. The most basic pieces of knowledge that are contained within this knowledge base is the first-level knowledge. These propositions are abbreviated using the first letter of each of its properties, so for example the Red Diamond Solitaire is represented as RDS.Įach player has a set of knowledge about the world, which we refer to as the knowledge base of the player. In this system, the most basic propositions are the Gem-Cards, where an occurrence of a Gem-Card indicates that the Gem-Card is in the game, and is thus not the hidden gem. The model that we created is based on S5 (m) logic, which is a mixed modal logic system. Players take turns in asking each other questions until one of the players can determine what the hidden gem is and subsequently win the game. ![]() After a question has been answered, the current player removes the used Question-Card from his hand and places back on the bottom of the deck, after which he draws a new Question-Card from the top of the deck. Apart from these two players, no other player can see this exchange of information. The latter player has to respond by showing all of his cards that match the properties of the Question-Card to the former player. In each round, the player whose turn it currently is can ask a question to one of the other players with the use of a Question-Card. There are 10 one-element, 33 two-element and 11 free-choice Question-Cards. The free-choice Question-Card can be used to ask either a one-element or a two-element question with properties that the player can choose him or herself. Red Solitaire), which players can use to ask for all of the Gem-Cards of another player that share both of these properties. Similarly, two-element Question-Card contain two properties (e.g. Red) and allows the player to ask another player for all of the Gem-Cards that he or she has that share that one property. A one-element Question-Card contains one property (e.g. There are three types of Question-Cards that the players can use to ask questions with: one-element, two-element and free-choice question cards. The deck of Gem-Cards consists of 36 unique cards that encompass all the possible combinations of Color (Red, Green, Yellow, Blue), Type (Diamond, Opal, Pearl) and Quantity (Solitaire, Pair, Cluster). Through logical deduction about the knowledge of the other players, players can discover what the hidden gem is and win the game. Each player is dealt four Question-Cards, which they can use on their turn to ask questions to other players about their cards. The rest of the cards are divided amongst the players. At the beginning of the game, one of these cards is removed from the deck: this is the hidden gem. The game consists of a deck of Gem-Cards, each with three properties: Color, Type and Quantity (e.g. Sleuth is a logical deduction game in which you need to reason about the knowledge of your opponents. The goal of this simulation is to show how the knowledge of the agents will lead to the decisions that they make and how they can use this information to alter their strategy. The model of the game is based on the S5 (m) logic system and the agents will use their knowledge about the world and the knowledge of other agents to make decisions and win the game. The application that we wrote can simulate a game of Sleuth with five AI Agents that play the game. This is a project for the Multi-Agent Systems course of 2016/2017, made by Richard Elderman & Cor Steging. This website contains information about a Simulator of the logical deduction card-game Sleuth. Sleuth Card Game Simulator Sleuth Card Game Simulator Multi-Agent Systems Course 2016-2017, University of Groningen View on GitHub Introduction
0 Comments
Leave a Reply. |