Stochastic Optimization of Area Navigation Noise Abatement Arrival and Approach Procedures

摘要

The use of stochastic kriging surrogate models for stochasticoptimization of area navigation arrivals and approaches has provedpromising for autonomously designing noise abatement procedures.The sample efficient algorithm exploits the dynamic structure of theproblem to quickly optimize new sample locations using acquisitionfunctions that iteratively explore and exploit the surrogate models.Reduction in the annoyance weighted sound energy of 40 to 50%,corresponding to expected reductions in the sound exposure level of2.4 to 3.1 dB, has been demonstrated in comparison to a conventionallydesigned RNAV procedure. Although the final approach, whichis the loudest portion of a procedure, was not considered in thisoptimization, the noise impact due to the preceding segments hasthe potential to be significantly diminished.