Top 5 Recent Articles
- Algorithms (19)
- All (400)
- Biography (1)
- Blog (45)
- Business Requirements (1)
- Commentary (1)
- Customers (2)
- Data Models (1)
- Education (2)
- GeoRaptor (13)
- GPS (1)
- Image Processing (2)
- Import Export (8)
- Licensing (2)
- LiDAR (1)
- Linear Referencing (4)
- Manifold GIS (3)
- Mapping (1)
- MySQL Spatial (7)
- Networking and Routing (including Optimization) (4)
- Open Source (18)
- Oracle Spatial and Locator (193)
- Partitioning (1)
- PostGIS (34)
- Published Articles (1)
- qGIS (1)
- Recommendations (1)
- Services (1)
- Software Change Log (1)
- Source Code (35)
- Space Curves (9)
- Spatial Database Functions (107)
- Spatial DB comparison (1)
- Spatial XML Processing (11)
- SQL Server Spatial (91)
- Standards (3)
- Stored Procedure (15)
- Tessellation or Gridding (10)
- Tools (2)
- Topological Relationships (1)
- Training (2)
- Triangulation (2)
There’s Network Analysis, and then there IS Network Analysis
Solutions are built by layering technology, from the simple to the complex.
Nothing more illustrates this than considering network analysis.
Firstly, most Geographic Information Systems (insert your favorite here), and database vendors like Oracle and PostgreSQL/PostGIS provide a basic routing engine.
And what they offer is generally basic: a network or graph model, functions for loading data such as street data into it, and some basic functions for querying such as Shortest Path and Traveling Salesman.
Sometimes one thinks there is a confusion about the basic offerings mentioned above: they are not a solution, only a basic tool.
Consider this: a basic network/routing package so mentioned, cannot provide functionality that can analyse a water network: there are too many domain-specific variables and functionality to implement and manage.
Specialist software is needed to develop and deliver solutions.
It has always been surprising that geospatial professionals confused their technologies with solutions.
Now, before Google Maps, there were precious little networking and routing solutions which were targeted at the public.
And Google Map has done this well: the ubiquitous network data; an excellent address search engine that includes not just street addresses but points of interest; business data; and street view.
Google Maps ability to generate directions from point A to point B with multiple stops is excellent.
But note: the stops have to be traversed in the input order: the stops are fixed.
What the user wants, the user wants!
What Google maps does well, it does well.
Users love it.
Note that, with Google, the only optimization is in the choice of streets between the requested points. However, the inputs to this are not just the street network but the time to travel a street determined by speed and capacity (e.g. dynamic load).
But when it comes to route optimization and scheduling of complex problems, e.g. OptimoRoute, real domain knowledge expressed in powerful algorithms and functionality is required.
Such functionality includes:
- User defined time and date constraints;
- Repeatable date schedules;/li>
- Inputs that include vehicle size (motor cycle through, van, to semi-rigid, and B-Double etc);/li>
- Loading and Unloading times;/li>
- Load priority;/li>
- Fuel Economy inputs./li>
As you can see there are lots of potential parameters that need to be implement; a basic network (graph) and routing engine isn’t enough.
In the past, real route optimization was specialist stuff only accessible by business that could afford it. But in these days of Cloud computer: Software As A Service (SAAS) etc, highly functional services can be accessed by businesses of any size at a price that fits their needs.
So, as was said at the start, the really interesting stuff, is the way technology is layered from the basic to the sophisticated.
A solution is available for most or all: just know what is needed and step out from there!