Energy-Efficient Task-Mapping for Data-Driven Sensor Network Macroprogramming Using Constraint Programming

Using constraint programming (CP), we address the taskmapping problem in data-driven macroprogramming for wireless sensor networks (WSNs). A task graph representing the flow of data among tasks assists the application developer in specifying the features of a WSN at a high level of abstraction. A problem that arises in this context is how to map the tasks to nodes in the target network before the deployment of sensors, in order to achieve an energy-efficient WSN. This problem is slightly different from the deployment problem for distributed systems. We take a published formulation of the WSN task-mapping problem solved by mixed integer programming (MIP) solvers, and rewrite it much more naturally as a constraint program, using off-the-shelf CP components. On realistic instances of real-world applications of the problem, we show that our CP model results in significantly better runtimes than the MIP model.