DYNAMIC PROGRAMMING MODEL FOR OPTIMAL LOCATION OF RUNWAY EXITS

Abstract The time required by a landing aircraft to clear the runway depends, among other things, of the type and location of runway exits available. For any given runway arrival pattern, in particular the aircraft separation times, the distribution of runway occupancy times determines the probability that the aircraft next in line for landing will be waved off. Horonjeff et al. (1959) prove that in the limit the expected runway acceptance rate is a function of the wave-off probability and then by the use of calculus determine optimal locations for up to three high-speed runway exits so as to maximize the expected runway acceptance rate for a bivariate normal distribution of runway deceleration distances and times. This note shows how this optimization can more efficiently be done by the use of dynamic programming for any arbitrary joint probability distribution of deceleration distances and times and any number of exits. The paper also explores several extensions to the basic model.