Synthetic Data Sets for Wireless Sensor Network Optimization


Nowadays, the technology of Wireless Sensor Networks (WSNs) is widely used. We can find examples of its use in many fields, such as environmental control, traffic control, precision agriculture, forest fire detection, rescue operations, among others. Sensors have some important features that promote the use of WSNs, such as they are small, powered by batteries, able to capture different physical variables in a same device, cheap, and so on. However, WSNs have important shortcomings that affect important aspects like quality of service and energy efficiency.

The efficiency deployment of WSNs has been defined as an NP-hard multiobjective optimization problem in the literature, so we can find several works trying to solve these inconveniences. These data sets are traditional WSNs (a set of sensors and a collector node) that we try to optimize adding routers through different strategies. This is the so-called Relay Node Placement Problem (RNPP). These data sets were successfully used in [1] and [2].

Data Set Properties

Sensors and sink node are placed on the same 2D-surface of size Dx x Dy. Sensors capture information packets from its environment with a sensibility radius Rs on a regular basis and simultaneously. Each time a sensor captures an information packet; this device must send it to the sink node following the routing protocol provided by Dijkstra’s algorithm for minimum distance among devices. The sink node is placed on the center of the scenario and collects all the information captured by the sensors.

Sensors are powered by batteries, and sink node has unlimited power supply. Each time a sensor sends an information packet, its battery suffers an energy cost, until its battery is exhaust, and then sensors cannot be used again. With the purpose of reducing the workload of sensors in communication task, a new device called router is added to WSNs. This device relays all the received information to the collector node. This way, sensors and collector node are provided by the data set, and routers coordinates are provided by the optimization algorithm. All devices (routers, collector node and sensors) can communicate among them with a communication radius Rc.

The instances are as follows:
     Instance 100x100_15_30. Scenario size: 100x100. Rc=30m. Rs=15m. 15 sensors.
     Instance 100x100_15_60. Scenario size: 100x100. Rc=60m. Rs=15m. 15 sensors.
     Instance 200x200_15_30. Scenario size: 200x200. Rc=30m. Rs=15m. 57 sensors.
     Instance 200x200_15_60. Scenario size: 200x200. Rc=60m. Rs=15m. 57 sensors.
     Instance 300x300_15_30. Scenario size: 300x300. Rc=30m. Rs=15m. 128 sensors.
     Instance 300x300_15_60. Scenario size: 300x300. Rc=60m. Rs=15m. 128 sensors.

Bellow, a zip file is provided with all the instance sets and some instructions. Zip file
If you used some of these instances, and you want to be referenced here, please contact us.

Related work

[1] "A New Realistic Approach for the Relay Node Placement Problem in Wireless Sensor Networks by Means of Evolutionary Computation". Jose M. Lanza-Gutierrez, Juan A. Gomez-Pulido, Miguel A. Vega-Rodríguez, Juan M. Sanchez-Perez. Ad Hoc and Sensor Wireless Networks, Old City Publishing, 2013, pag:-, ISSN:1551-9899. (Impact factor = 0.410-12).

[2] "Intelligent Relay Node Placement in Heterogeneous Wireless Sensor Networks for Energy Efficiency". Jose M. Lanza-Gutierrez, Juan A. Gomez-Pulido, Miguel A. Vega-Rodriguez and Juan M. Sanchez-Perez. International Journal of Robotics and Automation. Special Issue on Computational Intelligence and Sensor Networks for Automation Applications, Acta Press, 2013, pag:-, ISSN:0826-8185. (Impact factor = 0.494-12).

Contact us

Author: José Manuel Lanza Gutiérrez