Travelling Salesman Problem is a simple application build as the assignment of the "Developing Data Products" course conveyed by coursera.org.

The travelling salesman problem is a well-known mathematical problem that goes as follows:

There is a salesman that has a list of cities he needs to travel through and then return to the city of departure. The goal is to minimize the overall distance travelled.

The app allows user to choose between three different algorithms to calculate the shortest way:

• Nearest Neighbor
  
• Repetitive Nearest Neighbors
  
• Two edge exchange improvement procedure
  

The algorithms are embedded in the TSP package which is used in the application. Description of the algorithms can be found in the documentation.

1. Select places that the salesman needs to travel through by clicking on the map. Markers will pop-up with numeric IDs of the places. A list of the places will dynamically generate at the bottom-left part of the screen.
2. Select the Point of Departure.
3. Select the method
4. Click "Show the Way". A sequence of places will appear in the bottom-right corner. Also, the markers on the map will be updated by the visiting order.

The application does not use any geocoding.

It means that the places to travel are only identified by their latitude and longitude. No addresses, city names etc. are included.

It also means that the application does not take into account the actual roads. It only takes into account the shortest geographical distances.

Packages Used:

shiny, leaflet, TSP, geosphere

Links:

The application: https://sal2040.shinyapps.io/map_final/

Github repository with the code: https://github.com/Sal2040/Shiny—Travelling-Salesman