Network Analysis

From wiki.gis.com
Jump to: navigation, search
GIS Network Representations

Network Analysis in GIS is based on the mathematical sub-disciplines of graph theory and topology. Any network consists of a set of connected vertices and edges. Graph theory describes, measures, and compares graphs or networks. Topological properties of networks are: connectivity, adjacency, and incidence. These properties serve as a basis for analysis.[1] A simple example of a network in GIS can be streets, power lines, or city centerlines.

Overview of networks

GIS networks consist of interconnected lines (known as edges) and intersections (known as junctions) that represent routes upon which people, goods, etc. can travel.[2] The object traversing the network follows the edges, and junctions appear when at least two edges intersect.[3] Junctions and edges can have certain attributes affixed to them that increase the cost of traveling in the network, known as impedance. For example, a road network can have speed limits attached to the edges, and a junction can prevent left turns.[4]

Networks are either directed, in which only one direction of travel is allowed within the network, or undirected, in which any direction of travel is allowed.[5]

Types of networks that can be modeled in GIS

This shows a graphic of how the traveling salesman problem is displayed. The salesperson must find the shortest route that leads to each of the four points.

It's possible to perform analyses of movement within networks on GIS. These networks include:

  • Utility networks: including water mains, sewage lines, and electrical circuits. These networks are generally directed.
  • Transportation networks: including roads, railroads, and flight paths. These networks are generally undirected.[6]
  • Networks based on social connections.[citation needed]
The streets in this map were partitioned into zones based on the driving time to the nearest service point. No matter where you are in any of the zones, the point that is the shortest drive-time away is the one located within that zone. See: http://www.caliper.com/glossary/what-is-network-partitioning.htm

Shortest Path

One common type of network analysis is finding the shortest path between two points. In a network of streets, the "shortest" route can either refer to different variables, such as: distance, time, and monetary cost (such as purchasing a plane ticket). An ambulance driver looking for the shortest path to his destination will travel the route that will get him or her from point A to point B in the least amount of time.[7]


Traveling Salesman

The traveling salesman problem is defined as reaching every point in a network in the most efficient way possible. It is derived from the idea of a salesperson trying to reach a planned set of cities to sell his or her product in the quickest, most efficient way possible, either through money made, or time. UPS uses a traveling salesman algorithm to efficiently deliver as many packages as possible to their customers every day.[8]


Network Partition

Network partition is a divvying up of regions in a network to zones or subcategories.[9] These regions are sized based on proximity to specific points in a network. This is common for fire stations in metropolitan areas.


Transportation Modeling

Basic functions already existing in GIS including buffer, overlay, query, etc. are useful in Transportation Planning. However, deeper analysis of network data is available for planning applications. Examples of such higher uses include network flow equilibrium models, travel demand models, trip generation and distribution, as well as activity-based models and transportation/land-use interaction models. The latter use is particularly useful as demand for transportation influences land use, and reciprocally, the changed land-use's influence on transportation. Limited commercial software exists to perform such tasks, nevertheless, skilled programmers should be able design programs to perform these tasks where software is inaccessible or non-existent. See link #1 below for more information on Transportation and GIS.

Network Analysis Workflow

In order for you to perform a network analysis in a GIS program, there are basic steps you need to take [10]. This section walks through the general procedure to do before you solve network analysis network problems:

Organizing the Network Analysis Settings

In any GIS system, like ArcGIS, you need to enable the Network Analysis extension. You will also need to display the Network Analysis toolbar before you can perform any analysis.

Adding a Dataset to Your GIS Program

Before you are able to perform a network analysis, you need to create a dataset network in which you can perform an analysis. If there is not one available, the original reference source has been modified or has been changed, you will need to build a new one.

Creating the Network Analysis Layer

Layers contain an in-memory classes where inputs, properties, and results can be stored. [11] In the case of performing network analysis, the layer has to be connected to a network dataset. This layer has to be created and added to the dataset before the analysis can be performed. In ArcMap, a network dataset must be added first so that when an analysis layer is created, Network Analyst can bind the analysis layer to the network dataset. The six kinds of network analysis layers in ArcGIS include: route analysis layer, closest facility analysis layer, service area analysis layer, OD cost matrix analysis layer, vehicle routing problem analysis layer, location-allocation analysis layer. [12]

Imputing Network Analysis Features and Records

This step has us add features, or objects, to our dataset input. Network analysis objects are features and records used as input and output during network analysis. [13]These objects can include barriers, routes, facilities, or other man-made structures that will influence the end analysis. When you add these objects, setting the properties for the network analysis layer will further define the function of the input. [14]

Performing the Analysis

After you have finished the other steps of the procedure and created your layer, it is time to perform the network analysis. This can be done by clicking the Solve button on the Network Analysis toolbar you set up earlier in the process. Your results will then be displayed on the map and double-click the network analysis objects in the Network Analysis [15].

References

  1. Curtin, K., Network Analysis in Geographic information Science: Review, Assessment, and Projections Cartographic and Geographic Information Society
  2. Jensen, John R.; Jensen, Ryan R. (2013). Introductory Geographic Information Systems. Boston: Pearson. 195. 
  3. Jensen & Jensen, 201
  4. Jensen & Jensen, 202
  5. Jensen & Jensen, 200
  6. Jensen, John R.; Jensen, Ryan R. (2013). Introductory Geographic Information Systems. Boston: Pearson. 200. 
  7. What is Network analyst?
  8. http://geospatialrevolution.psu.edu/episode1
  9. http://www.caliper.com/glossary/what-is-network-partitioning.htm
  10. Network Analysis Workflow. ArcGIS for Desktop Extensions Help. ESRI. Retrieved from: http://desktop.arcgis.com/en/arcmap/latest/extensions/network-analyst/network-analysis-workflow.htm
  11. ESRI Network Analysis Workflow
  12. ESRI Network Analysis Workflow
  13. ESRI Network Analysis Workflow
  14. ESRI Network Analysis Workflow
  15. ESRI Network Analysis Workflow

Links

1. More on Transportation and GIS - https://people.hofstra.edu/geotrans/eng/methods/ch1m4en.html