Mathematical Assignment - open to bidding

Research goal: 1)My idea is to employ a machine learning technique to optimize the number of data streams for a given protocol in an adaptive manner to achieve the highest throughput. 2)How to improve the performance(mainly throughput) of the data transfer protocol. 3)How to Determine Optimal Dynamic Data Flow. 4)In a resource-sharing environment, which optimal dynamic data flow affects the host and the network. The planned steps of my research are: a) Run a given transfer protocol, vary the number of data streams, monitor host/network condition, and measure the corresponding transport performance. Note that TCP fairness may need to be considered for some UDP-based protocols if running on the internet. b) Select an appropriate machine learning algorithm(could start from a neural network), and train the model using the data collected at Step a) c) Build the model, measurement data, use Hadoop MapReduce / Spark for data analysis. Some existing machine learning modules in big data systems could be used. d) Once the model training is completed, apply the model for online testing.