Ongoing research project on data management and information retrieval within sensor networks under locally sensitive topology restrictions.
The instances utilized here come follows the following structure: the first line of the file gives out the total number of nodes within the network (modeled as a graph) as well as the sensor's communication range radius, followed by the sensors coordinates (x, y) separated from their respective neighbors list by an empty line (the first value corresponds to the total number of neighbors). An example of such instances is shown bellow, followed by its visual representation.
// Example graph
6 r
4 5.6
5.3 4.2
3.2 3.8
4 2
1 3
2 1
2 1 2
1 0
3 0 3 4
2 2 5
2 2 5
2 3 4
This work considers the case where all the sensors within the network have the same communication area. Sensor nodes p and q are said to be neighbors (e.g. vertices have an edge connecting them) if their communication range overlap.
- With Acks
- Without Acks
The code follows an Object Oriented programming paradigm where each node of is represented as an object of a class that encapsulates the sensor, where the knowledge of a sensor u (local sensibility) about the network topology is limited to its neighboring sensors listed on the instance file.
In order to simulate a network of communicating and independet sensors the program utilizes a message sharing interface between processes, where each processe represents a sensor of the network.
The source code files can be found here along with instructions on needed dependencies as well as how to compile and run.