Skip to content

mtresnik/graph

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

24 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

graph

build status version License PRs Welcome

A novel Graph-Theory and Maze Solution library made for kotlin JVM.


Dependencies

Please compile against com.resnik.math:1.0.0 and com.resnik.intel:1.0.0 as well as this project.

Getting Started

This is a slightly different process to that of com.resnik.intel.

Maven

~/.m2/settings.xml:

<settings xmlns="http://maven.apache.org/SETTINGS/1.0.0"
  xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
  xsi:schemaLocation="http://maven.apache.org/SETTINGS/1.0.0
                      http://maven.apache.org/xsd/settings-1.0.0.xsd">
    ...
  <activeProfiles>
    <activeProfile>github</activeProfile>
  </activeProfiles>
    ...
  <servers>
    <server>
      <id>github</id>
      <username>GITHUB_USERNAME</username>
      <password>GITHUB_PAT</password>
    </server>
  </servers>
</settings>

pom.xml:

<repository>
    <id>github</id>
    <url>https://maven.pkg.github.com/mtresnik/math</url>
    <snapshots>
        <enabled>true</enabled>
    </snapshots>
</repository>

<repository>
    <id>github</id>
    <url>https://maven.pkg.github.com/mtresnik/intel</url>
    <snapshots>
        <enabled>true</enabled>
    </snapshots>
</repository>

<repository>
    <id>github</id>
    <url>https://maven.pkg.github.com/mtresnik/graph</url>
    <snapshots>
        <enabled>true</enabled>
    </snapshots>
</repository>
...
<dependency>
    <groupId>com.resnik</groupId>
    <artifactId>math</artifactId>
    <version>1.0.0</version>
</dependency>

<dependency>
    <groupId>com.resnik</groupId>
    <artifactId>intel</artifactId>
    <version>1.0.0</version>
</dependency>

<dependency>
    <groupId>com.resnik</groupId>
    <artifactId>graph</artifactId>
    <version>1.0.0</version>
</dependency>

Gradle (groovy)

~/.gradle/gradle.properties:

gpr.user=GITHUB_USERNAME
gpr.token=GITHUB_PAT

build.gradle:

repositories {
    ...
    maven {
        url= uri("https://maven.pkg.github.com/mtresnik/math")
        credentials {
            // Runner stored in env, else stored in ~/.gradle/gradle.properties
            username = System.getenv("USERNAME") ?: findProperty("gpr.user") ?: "<GITHUB_USERNAME>"
            password = System.getenv("TOKEN") ?: findProperty("gpr.token")
        }
    }
    ...
    maven {
        url= uri("https://maven.pkg.github.com/mtresnik/intel")
        credentials {
            // Runner stored in env, else stored in ~/.gradle/gradle.properties
            username = System.getenv("USERNAME") ?: findProperty("gpr.user") ?: "<GITHUB_USERNAME>"
            password = System.getenv("TOKEN") ?: findProperty("gpr.token")
        }
    }
    maven {
        url= uri("https://maven.pkg.github.com/mtresnik/graph")
        credentials {
            // Runner stored in env, else stored in ~/.gradle/gradle.properties
            username = System.getenv("USERNAME") ?: findProperty("gpr.user") ?: "<GITHUB_USERNAME>"
            password = System.getenv("TOKEN") ?: findProperty("gpr.token")
        }
    }
}

dependencies {
    ...
    implementation group: 'com.resnik', name: 'math', version: '1.0.0'
    implementation group: 'com.resnik', name: 'intel', version: '1.0.0'
    implementation group: 'com.resnik', name: 'graph', version: '1.0.0'
    ...
}

Mazes

Maze Generation

Minimum Spanning Trees (MST's)

Prim's Algorithm (40x40) Kruskal's Algorithm (40x40)

The obvious benefit to MST's is that every node can be reached from any other node, meaning all produced mazes are consistent (able to be solved).

"Recursive" Subdivision (40x40) Aldous-Broder Algorithm

Recursive is in quotes here because the actual process of generating the Maze uses Depth First Search in a single method rather than programatically calling itself.

The Aldous-Broder Algorithm is slightly modified such that neighboring frontiers are connected. This produces a more uniform spanning tree.

WIP : Recursive DFS for Maze Generation

Maze To Graph Conversions

Initial Maze Graph Representation
val maze : Maze = // Somehow generated maze...
val graph = MazeToGraphProvider(maze).build()

The resulting graph represents the initial maze, where a MazeCell is represented by a Vertex and a MazeBorder determines the Edge's.

Maze Solution Graph Solution

All GraphAlgorithms can be used on Graphs and all Mazes can be converted into Graphs.

Traversals

Path Finding

  • Breadth-First Search
  • Depth-First Search
  • Dijkstra
  • A* Search

Minimum Spanning Trees (MST)

Kruskal's Algorithm Prim's Algorithm

Generated using PartiallyConnectedGraph with V=20 and (E/V)<=10

In general, the MST algorithms accept a base Graph and return a cloned Graph representing the Tree. (Formally this is written as G \ T)

Traveling Salesperson Problem (TSP)

  • Brute Force Search - O(n!) uses recursion
  • Permutation Search - O(n!) linear search
  • Random Search - O(N * (|V| + |E|))
  • Greedy Search (sub optimal) - O(|V| + |E|)
  • Greedy Twice-Around (uses Prim's MST) - O(|V|^2 + MST)
  • Two-Opt (reduces edge cross over)

Serialization

Vertices, Edges, Paths, and Graphs are both Clonable and Serializable.

Saving an Identifyable Item to an ItemizedLongStorable will set an auto-incementing ID (long) to the Item and store it within its internal collection. In general, there is VertexStorage, EdgeStorage, PathStorage, and GraphStorage.

/*
 * Where xyzStorage could be: 
 * VertexStorage, 
 * EdgeStorage, 
 * PathStorage, or GraphStorage
 * */
val outputStream = ByteArrayOutputStream()
xyzStorage.writeTo(outputStream)

// or to file...
val parent = File("PATH/TO/GRAPH/STORAGE")
xyzStorage.saveFromParent(parent)

VertexStorage

Usage

val vertexStorage = graph.storage.vertexStorage
// or
val vertexStorage2 = VertexStorage()
vertexStorage2.save(v1)
vertexStorage2.save(v2)
// ...
vertexStorage2.save(vn)

vertexStorageX.forEach{ vertex -> doSomething(vertex) }
header v size 3 
header v bbox 0.1 1.0 11.0 2.0 
v 1 0.5 1.0 | 1 5 11
v 2 0.1 2.0
v 3 11.0 2.0 | 12

EdgeStorage

Usage

val edgeStorage = graph.storage.edgeStorage
// or
val vertexStorage = VertexStorage()
val edgeStorage2 = EdgeStorage(vertexStorage)
vertexStorage.save(v1)
vertexStorage.save(v2)
val edge1 = Edge(v1, v2)
edgeStorage2.save(edge1)

edgeStorageX.forEach { edge -> doSomething(edge) }
header e size 3 
e 1 1 2 0.0
e 2 2 3 0.0
e 3 3 1 0.0

GraphStorage

header g v PATH\TO\VERTICES\file.rgv
header g e PATH\TO\EDGES\file.rge 
header g t PATH\TO\PATHS\file.rgt 

Graph Providers

val graphProvider = BoundedGraphProvider(bbox, width, height)
val graph = graphProvider.build()

// Prune input by 20%
val prunedProvider = RandomPruneGraphProvider(graph, 0.20)

// Uses a cloned graph for pruning
val pruned1 = prunedProvider.build()
val pruned2 = prunedProvider.build()
// ...
val prunedN = prunedProvider.build()

Nearest Neighbor

val vertexStorage = graph.storage.vertexStorage

// O(|V|)
val start = vertexStorage.nearestNeighbor(ArrayPoint(0.5, 1.0))
val dest = vertexStorage.nearestNeighbor(ArrayPoint(10.5, -20))

// kNN checks : O(k*|V|)
val closest3 = vertexStorage.kNearestNeighbors(ArrayPoint(3.0, 4.0), 6)