No abstract

No abstract

No abstract

Scheduling tasks in a multiprocessor system is found to be a NP-hard problem and a considerable amount of time is used up when it is solved using conventional techniques. Therefore, evolutionary algorithms like Genetic Algorithms (GA) have been explored for scheduling tasks in a multiprocessor system. GA can be implemented in various manners. This paper investigates the performance of GA with two different selection operators. This paper also studies how introducing elitism effects the performance of GA. Extensive simulations have been carried out in order to find the better candidate among the two selection operators. The decision is made depending on stability of the GA output, the rate of convergence of output and the ability of GA to give an output which is as close as possible to the actual output.

Summary Perpetual lifetimes and low maintenance are few of the attractive aspects of sensor nodes that harvest ambient energy. However, their operation depends heavily on the energy profiles of their harvesting source(s). In this work, we study the suitability of energy harvesting sensor (EHS) Nodes, powered using indoor lighting and vibrations, for a simple temperature monitoring application. To help these nodes sustain, we have proposed schemes that allow the nodes to sample and transmit data judiciously. Along with an adaptive sampling based on autoregressive (AR) model, we have proposed a regulating function based transmission scheme that regulates the amount of transmitted data based on the energy available at the node and the characteristics of the data so that even with limited transmissions the fidelity of the data is not lost. With the help of thorough evaluations, we can conclude that an EHS node fares quite well. Results show that adaptive sampling and transmissions based on regulating function not only save energy at the sensor nodes, but they also reduce the amount of data generated and accumulated in a network.