This is the repository for the 2019 SSRP; making an AI for Eurorails.

This project uses Gradle as its build system. To run anything, you need the Java 8 JDK installed.

Running the game can be done with gradlew desktop:run. This launches the game with no options. Note that there is no AI in this mode.

To enable an option, add it after the run command. For example, running gradlew desktop:run -Pdebug enables debug mode.

Note that these options also exist when running the built jar file, but without the P prefix. To enable debug mode, for example, you could run java -jar desktop-release.jar -debug.

To build an output jar file in build/libs/desktop-1.0.jar: gradlew desktop:dist

Core source code is located in core/src/empire.

• Code for the AI is in the ai package. CurrentAI is the core AI class.
• Code for the game itself is located in game. State handles game state and has most of the control methods.
• Code for graphical effects and rendering is in gfx.
• Code for networking is in net, and code for reading data input files is in io.

All input files and other assets are in core/assets/. Raw unpacked files are available in core/assets-raw/ if you need them. Run ./gradlew pack to pack these sprites into a spritesheet.

Data for the map is in core/assets/maps/eurorails.txt. Text data for the deck is in core/assets/maps/deck.txt. While the card data should be self explanatory, you can see how it is parse in core/src/empire/io/CardIO.java. Code for parsing the map can be seen in core/src/empire/io/MapIO.java.

All the code for controlling AI is located in core/src/empire/ai/CurrentAI.java. While most of the code should already be documented, there are a few methods of special importance that may need to be explained:

• #act() runs every frame, and is responsible for updating all the AI's systems and performing moves.
  • If there is currently no plan and nothing being calculated, the AI launches a task to update the plan in an new thread.
  • If a plan is currently being calculated, it does nothing.
  • If a plan is available, the AI attempts to execute it.
• #executePlan() handles the actual execution of the plan, going through each action and attempting to perform it.
  • This handles upgrading the locomotive, linking cities, moving, placing track, loading/unloading, etc.
  • Actions outside the plan may also be executed (e.g. loading up random cargo just in case it happens to get a card for it).
• #updatePlan() is a blocking method to find the best plan and set it up.
  • This method goes through all combinations of plans, then substitutes them for combinations of actions (load/unload).
  • Each plan is evaluated and the best one (smallest cost) is selected.
• Plan#cost() evaluates the cost of a plan.
  • This is done by adding up the total money spent placing track, multiplying it by some number and subtracting the total profit * some multiplier
  • The "profit multiplier" is CurrentAI#demandCostScale.
• Cost of placing track between locations is calculated in ai/Astar.java.
  • This is measured in movement points.
  • Cost of moving on already-placed track is 1 movement point
  • Cost of placing track/spending money is 1 + (cost) * multiplier, where multiplier is defined in Astar#costScale() - currently 6.
• To get the final cost of a plan, the total profit multiplied by the profit multiplier is subtracted from the A* path cost.
  • The plan of least cost is picked.

Each actions the player can do is represented as a class. These classes are located in game/Actions. Each action must implement the apply method, which executes the action. Do not use this method directly. This is designed for easy use in multiplayer. PlayerAction is an abstract class for actions that are specific to a single player, e.g. moving or placing track.

To execute an action, simply do:

new TypeOfAction(){{ //replace TypeOfAction with whatever action you want to perform
    //set up parameters here, e.g. track placed, tile moved to
}}.act(); //calling act() executes it

This project uses websockets for multiplayer. A server is hosted on the port specified in net/Net::port when you press "host." To connect, simply enter the IP of the host machine and press 'connect.'