Chapter 14

Maze Game

This chapter examines how to develop a maze-based game, where the character is to collect all the coins in a maze whilst avoiding a ghost. The chapter expositions algorithm on how to generate random mazes using a depth-first algorithm, as well as finding the shortest path to the character from the ghost using a breadth-first algorithm

As suggested by the book the following features have been added:

• Added a menu screen
• Added more audio effects
• Added a timer to the UI
• Increasing ghost speed with time
• Hero health points
• Added more ghosts
• Implemented a larger maze

Bonus features:

• camera/character alignment
• camera zoom
• camera boundaries with zoom

Maze Generation

Instead of using Tiled to create several maps a unique map may be generated each time a new level starts. Such procedural content generation greatly adds to replayability.

The general undertaking is as follows:

1. Create a rectangular grid of rooms, all surrounded with walls
2. Remove walls between rooms such that every room may be reached.
   1. Select a room as a starting location; mark it as connected and add it to the list
   2. While there are still rooms remaining in the list:
      1. Let currentRoom be the most recently added room from the list
      2. If currentRoom has any unconnected neighbors
         1. let nextRoom be a random unconnected neighbor of currentRoom
         2. remove the walls between currentRoom and nextRoom
         3. mark nextRoom as connected; and
         4. add nextRoom to the end of the list
      3. If currentRoom does not have any unconnected neighbors, remove it from the list
3. Remove additional random walls to create multiple paths to rooms.

This depth-first algorithm creates a random path through the maze for as long as possible and then starts anew.

Pathfinding

The ghost follows the shortest path to the character. All rooms that are one square away will be checked first, and so forth. This is a breadth-first search algorithm. Pseudocode for the algorithm:

1. Identify startRoom and targetRoom.
2. Set currentRoom equal to startRoom.
3. Mark currentRoom as visited, set its previous room to null, and add currentRoom to a list.
4. While the list is not empty, do the following:
   1. Set currentRoomto the first element in the list and remove it from the list.
   2. For each unvisited neighbor (called nextRoom) of currentRoom,
      1. set nextRoom's previous room to currentRoom.
      2. end the algorithm if nextRoom is targetRoom; and
      3. if nextRoom is not targetRoom, then mark nextRoom as visited and add nextRoomto the end of the list.

Camera

A new camera functionality was added to BaseActor, the ability to zoom:

fun zoomCamera(zoom: Float) {
    if (this.stage != null) {
        val camera = this.stage.camera as OrthographicCamera
        camera.zoom = zoom

        bindCameraToWorld(camera)
        camera.update()
    }
}

This required an update to the bindCameraToWorld function, as it now also must take into account the zoom level:

private fun bindCameraToWorld(camera: OrthographicCamera) {
    camera.position.x = MathUtils.clamp(
        camera.position.x,
        (camera.viewportWidth * camera.zoom) / 2,
        worldBounds.width - (camera.viewportWidth * .9f) / 2
    )
    camera.position.y = MathUtils.clamp(
        camera.position.y,
        (camera.viewportHeight * camera.zoom) / 2,
        worldBounds.height - (camera.viewportHeight * .9f) / 2
    )
}

New Imports

No new imports

Previous | Next