Planning the most efficient route with Optimap

Does your organization provide services in the community regularly?  Do you have regular routes of sites to visit: neighborhood elderly, community centers, or even regular errand runs?  There are many situations when a local nonprofit or community organization needs to plan a logical route between multiple locations in a city or region.  But what's the most efficient route when you have several - or a dozen - sites to visit?

In the computer science world, this is a classic computational challenge that programmers have long studied, referred to as the "traveling salesman problem" because of its original context: given the distances between each city, what is the most efficient route between a number of cities, so that you visit each city once and only once?

There are lots of community organizations that need effective route-planning software, especially with transportation and personnel costs rising: shuttle buses for children and other non-driving community members, "Meals on Wheels" programs for the elderly or ill, physical donation pickups and deliveries, and more.  Thankfully, there are free easy-to-use online services for computing these optimal routes between locations - today we'll take a look at one created and offered for free by a Norweigian programmer, Optimap. 

Optimap allows users to enter a number of locations and then generates the optimal route between all of the locations, using Google's geocoding and routing service, displaying the results on a Google Map.  The accompanying street directions can be printed out or downloaded to a TomTom GPS navigation device.  Users can choose to select surface roads, avoiding highways - or even avoid roads altogether by getting pedestrian directions.

Users can add locations to the planner by clicking on a Google map, entering a list of addresses in a text box, or entering a list of pre-calculated latitude and longitude coordinates.  Although users can enter up to 24 locations to be routed, the "optimal" trip is only planned for up to 15 locations.  This limitation is due to the fact that the "traveling salesman problem" is computationally intensive, even today.