Communication scheduling in data gathering networks of heterogeneous sensors with data compression: Algorithms and empirical experiments

We consider a communication scheduling problem to address data compression and data communication together, arising from the data gathering wireless sensor networks with data compression. In the problem, the deployed sensors are heterogeneous, in that the data compression ratios, in terms of size reduction, the compression time, and the compression costs, in terms of energy consumption, on different sensors are different. The bi-objective is to minimize the total compression cost and to minimize the total time to transfer all the data to the base station. The problem reduces to two mono-objective optimization problems in two separate ways: in the original problem a time bound is given and the mono-objective is to minimize the total compression cost, and in the complementary problem a global compression budget is given and the mono-objective is to minimize the makespan. We present a unified exact algorithm for both of them based on dynamic programming; this exact algorithm is then developed into a fully polynomial time approximation scheme for the complementary problem, and a dual fully polynomial time approximation scheme for the original problem. All these approximation algorithms have been implemented and extensive computational experiments show that they run fast and return the optimal solutions almost all the time.