曝光台 注意防骗
网曝天猫店富美金盛家居专营店坑蒙拐骗欺诈消费者
The initial position of A1 is (0, 50000) and FL100. The approach maneuver of this aircraft is fixed to ω¯1 = 30000 and ω¯2 = 50000. The initial position of A2 is (0, 50000) and altitude FL100. The parameters of its approach maneuver will be selected using the optimization algorithm. The range of the optimization parameters is ω2 ∀ [35000, 60000] and ω1 ∀ [0,ω2]. We assume that the performance of the approach maneuver is measured by the arrival time of A2 at the start of the glide path (T2). The measure of performance
−a·T2
is given by perf = ewith a =5 · 10−4 . The constraint is that the trajectories of the two aircraft are not in conflict. In addition, aircraft 2 must also reach the altitude of 1500 ft before the start of the glide path, to ensure that it intercepts the localizer. We optimize
¯
initially with an upper bound on probability of constraint violation given by P =0.1. Since there exists a maneuver in the set of optimization parameters that gives negligible conflict probability, we select A = 10 in the optimization criterion according to inequality (7).
The results of the optimization procedure are illustrated in Figures 6(b-d) for different values of J and proposal distribution g. Each figure shows the scatter plot of the accepted parameters during MCMC simulation. Again, the first 10% of accepted parameters were discarded in each case as a burn in period. Figure 6(b) illustrates the case J = 10 and proposal distribution g uniform over the parameter space. In this case, the ratio
14
[feet] [feet]
Positions after 10 minutes
10000
9000
8000
7000
6000
5000
4000
3000
2000
1000
0
6 5 4 3 0 −12 −2
1 −3
x 10
4
−4
0 −5 −1
−6
4
x 10
[meters]
[meters]
Positions after 20 minutes
10000
9000
8000
7000
6000
5000
4000
3000
2000
1000
0
6
5
4
3 0−12 −2
1 −3
x 10
4
−4
0 −5 −1
−6
4
x 10
[meters] [meters]
Figure 5: Several trajectory realizations of an approach maneuver
between accepted and proposed parameters during MCMC simulation was 0.23. This means that approximately 1100 · 10/0.23 = 47826 simulations were needed to obtain 1000 accepted states, which took approximately 2.6 hours. A region characterized by a low density of accepted parameters can be clearly seen in the figure. These are parameters which correspond to a conflicting maneuver where the aircraft are performing an almost symmetrical approach. The figure also shows two distinct “clouds” of accepted maneuvers. They correspond to a discrete choice that the air traffic controller has to make: either land A2 before A1 (bottom right cloud) or land A1 before A2 (top left cloud).
中国航空网 www.aero.cn
航空翻译 www.aviation.cn
本文链接地址：
Monte Carlo Optimization for Conflict Resolution in Air Traffic Control(13)
