Example usage of clean_network in GAMA Platform

Introduction

The clean_network operator can be used used for cleaning and optimizing road networks. It helps deal with issues in GIS data such as disconnected segments, overlapping roads, and imprecise geometry.

Syntax

clean_network(
    list<geometry> geometries,
    float tolerance,
    bool split_lines,
    bool keepMainConnectedComponent
)

Parameters Explained

1. geometries (list)

• What it is: Collection of geometries (usually road segments) to clean
• How to get it: road collect each.shape or road_shapefile.contents
• Example:

list<geometry> road_geometries <- road collect each.shape;
list<geometry> road_geometries <- road_shapefile.contents;

2. tolerance (float)

• What it is: Maximum distance between points to be considered for merging
• Recommended values:
  • Small networks: 0.1 to 1.0 (e.g., urban network)
  • Large networks: 1.0 to 5.0 (e.g., rural network, highway network)
• Effects:
  • Lower values (e.g., 0.1): More precise, less merging
  • Higher values (e.g., 5.0): More aggressive merging
• Example scenarios:
  • For precise road networks: tolerance <- 0.5
  • For rough networks: tolerance <- 2.0

3. split_lines (boolean)

• What it is: Whether to split roads at intersection points
• Values:
  • true: Split roads at intersections (recommended for navigation)
  • false: Keep roads as continuous lines
• When to use:
  • true: For routing and navigation purposes
  • false: For visualization or analysis only

4. keepMainConnectedComponent (boolean)

• What it is: Whether to keep only the main connected network
• Values:
  • true: Remove isolated road segments
  • false: Keep all road segments
• When to use:
  • true: For ensuring network connectivity
  • false: When isolated segments are important

Common Usage

Basic Network Cleaning

// Basic cleaning with default values
list<geometry> clean_lines <- clean_network(
    road collect each.shape,
    1.0,    // moderate tolerance
    true,   // split at intersections
    true    // keep only main network 
);

Implementation Example

global {
  // Network cleaning parameters
  float tolerance <- 1.0;
  bool split_lines <- true;
  bool keep_main_connected <- true;

    init {
        // Create initial roads
        create road from: road_shapefile;
        
        // Clean the network
        list<geometry> clean_lines <- clean_network(
            road collect each.shape,
            tolerance,
            split_lines,
            keep_main_connected
        );
        
        // Remove old roads
        ask road {
            do die;
        }
        
        // Create new roads
        loop clean_geom over: clean_lines {
            create road {
                shape <- clean_geom;
            }
        }
        
        // Create network graph
        road_network <- as_edge_graph(road);
    }
}