Multi-agent algorithms for adaptation of underwater acoustic communication parameters

In this work I studied the problem of adaptation of the communication parameters for orthogonal frequency-division multiplexing communication schemes through a constrained distributed convex-optimization with asyn- chronous computation and lossy communication. The algorithm implemented was the so called robust-asynchronous Newton-Raphson Consensus while the constraints were added using an interior point method with logarithmic bar- riers. Two other algorithms such as Gradient Descent and Jacobi Descent were also tested in the work and compared to the main one. Regarding the cost function, I proposed a maximization of the throughput per OFDM packet without having the packet loss ratio exceed a predefined threshold, to improve the performance of the communication for the whole network of agents while automatically adjusting the channel’s parameters along with possible different sea environmental conditions. Numerical simulations shows the results obtained and the performance that this solution can provide to the acoustic communication schemes.