Multiobjective Departure Runway Scheduling Using Dynamic Programming

Abstract

At busy airports, air traffic controllers seek to find schedules for aircraft at the runway that aim to minimize delays of the aircraft while maximizing runway throughput. In reality, finding optimal schedules by a human controller is hard to accomplish since the number of feasible schedules available for the scheduling problem is quite large. In this paper, we pose this problem as a multiobjective optimization problem, with respect to total aircraft delay and runway throughput. Using principles of multiobjective dynamic programming, we develop an algorithm to find a set of Pareto-optimal solutions that completely specify the nondominated frontier. In addition to finding these solutions, this paper provides a proof of the algorithm's correctness and gives an analysis of its performance against a baseline algorithm using the operational data for a model of the Dallas/Fort Worth International Airport.Index Terms-Aircraft scheduling, multiple objective dynamic programming, Pareto-optimality, runway scheduling.

Distance Evaluated Simulated Kalman Filter with State Encoding for Combinatorial Optimization Problems

“…During the initialization of agents, a random bit string, , is generated for each agent. During the generation of solution, the term in (5) is mapped into a probabilistic value in the range of [0,1] using a mapping function, , as follows: (7) where . Then, the probabilistic value is compared to a random number, rand , to update the solution If the random number is smaller than the probabilistic value, the binary number at that particular dimension is complemented.…”

“…Examples of numerical optimization problems are Proportional-Integral-Derivative (PID) tuning in control applications [1], training of a neural network [2], adaptive beamforming [3], and engineering design problems [4]. On the other hand, combinatorial optimization or discrete problems concern with the best combination of a set of variables such as in traveling salesman problems [5], assembly sequence planning [6], scheduling for airport [7] and classroom [8], routing in integrated circuits [9], and feature selection [10]. In solving discrete and combinatorial optimization problems, metaheuristic algorithms such genetic algorithm (GA) [11] and ant colony optimization (ACO) [12] have been developed to operate in binary search space.…”

Optimization of Airport Surface Traffic: A Case-study of Incheon International Airport

“…The runway scheduling problem has been studied in various ways. In particular, many common solution techniques used for job-shop scheduling problems, such as dynamic programming 9,11,20,22 , heuristics 24 , as well as Mixed Integer Linear Programming (MILP)-based optimization model 3,4,13,8,10 , were applied to runway scheduling problems. Ref.…”

Exact and Heuristic Algorithms for Runway Scheduling

“…There are many dynamic programming formulations for various types of runway scheduling. [8][9][10][11][12][13] The proposed algorithm uses a state definition similar to the one in Ref. [8] and [10] and adds additional information to the state to handle the constraints specific to SRS.…”

“…A common objective is to minimize the completion time (runway-use time) of the last job, which is equivalent to maximizing throughput. Hence, many of the solution techniques commonly used for solving the job shop scheduling problems have been adapted to the runway scheduling problem; e.g., mixed integer linear programs, [5][6] branch and bound, 7 dynamic programming, [8][9][10][11][12][13] heuristics, [14][15][16] metaheuristics, [17][18] and others. The sequence dependent job shop scheduling problem is strongly NPhard, and consequently, it is not expected to find polynomial time algorithms for the runway scheduling problem.…”