Cover -- Title Page -- Copyright -- Contents -- Introduction -- Chapter 1: The Context of Air Traffic Management -- 1.1. Introduction -- 1.2. Vocabulary and units -- 1.3. Missions and actors of the air traffic management system -- 1.4. Visual flight rules and instrumental flight rules -- 1.5. Airspace classes -- 1.6. Airspace organization and management -- 1.6.1. Flight information regions and functional airspace blocks -- 1.6.2. Lower and upper airspace -- 1.6.3. Controlled airspace: en route, approach or airport control -- 1.6.4. Air route network and airspace sectoring -- 1.7. Traffic separation -- 1.7.1. Separation standard, loss of separation -- 1.7.2. Conflict detection and resolution -- 1.7.3. The distribution of tasks among controllers -- 1.7.4. The controller tools -- 1.8. Traffic regulation -- 1.8.1. Capacity and demand -- 1.8.2. Workload and air traffic control complexity -- 1.9. Airspace management in en route air traffic control centers -- 1.9.1. Operating air traffic control sectors in real time -- 1.9.2. Anticipating sector openings (France and Europe) -- 1.10. Air traffic flow management -- 1.11. Research in air traffic management -- 1.11.1. The international context -- 1.11.2. Research topics -- Chapter 2: Air Route Optimization -- 2.1. Introduction -- 2.2. 2D-route network -- 2.2.1. Optimal positioning of nodes and edges using geometric algorithms -- 2.2.2. Node positioning, with fixed topology, using a simulated annealing or a particle swarm optimization algorithm -- 2.2.3. Defining 2D-corridors with a clustering method and a genetic algorithm -- 2.3. A network of separate 3D-tubes for the main traffic flows -- 2.3.1. A simplified 3D-trajectory model -- 2.3.1.1. 3D-trajectories with lateral or vertical deviations -- 2.3.1.2. Deviation costs for the simplified model -- 2.3.1.3. A simple criterion for 3D-trajectory separation.

2.3.2. Problem formulations and possible strategies -- 2.3.2.1. Sequential approach or global optimization -- 2.3.2.2. Problem difficulty and choice of algorithms -- 2.3.3. An A* algorithm for the "1 versus n" problem -- 2.3.3.1. General description of tree-search or graph-search methods -- 2.3.3.2. The A* algorithm -- 2.3.3.3. State space representation for the simplified 3D-trajectory model -- 2.3.3.4. Cost and heuristic for the simplified 3D-trajectory model -- 2.3.4. A hybrid evolutionary algorithm for the global problem -- 2.3.4.1. General description of an evolutionary algorithm -- 2.3.4.2. Encoding individuals -- 2.3.4.3. Initial population -- 2.3.4.4. Fitness criterion -- 2.3.4.5. Parent selection -- 2.3.4.6. The crossover operator -- 2.3.4.7. The mutation operator, for simple test-cases -- 2.3.4.8. The mutation operator for complex test-cases and real data -- 2.3.4.9. Selecting individuals for the new population -- 2.3.5. Results on a toy problem, with the simplified 3D-trajectory model -- 2.3.5.1. Description of the toy problem and test-cases -- 2.3.5.2. Results of the A* algorithm on the toy problem -- 2.3.5.3. Results of the evolutionary algorithm on the toy problem -- 2.3.6. Application to real data, using a more realistic 3D-tube model -- 2.3.6.1. Traffic flow model and computation -- 2.3.6.2. A more realistic 3D-tube model based on aircraft performances -- 2.3.6.3. Attributes of the 3D-tubes -- 2.3.6.4. Detection of 3D-tubes intersections -- 2.3.6.5. Adaptation of the algorithms -- 2.3.6.6. Results in the French airspace -- 2.3.6.7. Results in the European airspace -- 2.4. Conclusion on air route optimization -- Chapter 3: Airspace Management -- 3.1. Airspace sector design -- 3.2. Functional airspace block definition -- 3.2.1. Simulated annealing algorithm -- 3.2.2. Ant colony algorithm -- 3.2.3. A fusion-fission method.

3.2.4. Comparison of fusion-fission and classical graph partitioning methods -- 3.3. Prediction of air traffic control sector openings -- 3.3.1. Problem difficulty and possible approaches -- 3.3.2. Using a genetic algorithm -- 3.3.3. Tree-search methods, constraint programming -- 3.3.4. A neural network for workload prediction -- 3.3.5. Conclusion on the prediction of sector openings -- Chapter 4: Departure Slot Allocation -- 4.1. Introduction -- 4.2. Context and related works -- 4.2.1. Ground holding -- 4.2.1.1. Satisfying sector capacity constraints -- 4.2.1.2. Solving the conflicts -- 4.3. Conflict-free slot allocation -- 4.3.1. Conflict detection -- 4.3.2. Sliding forecast time window -- 4.3.3. Evolutionary algorithm -- 4.3.3.1. Variables and data structures -- 4.3.3.2. Constraints -- 4.3.3.3. Fitness function -- 4.3.3.4. Mutation operator -- 4.3.3.5. Crossover operator -- 4.3.3.6. Sharing -- 4.3.3.7. Parameters -- 4.4. Results -- 4.4.1. Evolution of the problem size -- 4.4.2. Numerical results -- 4.5. Concluding remarks -- Chapter 5: Airport Traffic Management -- 5.1. Introduction -- 5.1.1. Airports' main challenges -- 5.1.2. Known difficulties -- 5.1.3. Optimization problems in airport traffic management -- 5.2. Gate assignment -- 5.2.1. Problem description -- 5.2.2. Resolution methods -- 5.3. Runway scheduling -- 5.3.1. Problem description -- 5.3.2. An example of problem formulation -- 5.3.3. Resolution methods -- 5.4. Surface routing -- 5.4.1. Problem description -- 5.4.2. Related work -- 5.5. Global airport traffic optimization -- 5.5.1. Problem description -- 5.5.2. Coordination scheme between the different predictive systems -- 5.5.3. Simulation results -- 5.6. Conclusion -- Chapter 6: Conflict Detection and Resolution -- 6.1. Introduction -- 6.2. Conflict resolution complexity -- 6.3. Free-flight approaches -- 6.3.1. Reactive techniques.

6.3.2. Iterative approach -- 6.3.3. An example of reactive approach: neural network trained by evolutionary algorithms -- 6.3.3.1. Problem modeling -- 6.3.3.2. The inputs -- 6.3.3.3. The neural network structure -- 6.3.3.4. Learning the neural network weights -- 6.3.3.5. Evolutionary algorithm used -- 6.3.3.6. Computing the fitness -- 6.3.3.7. The learning examples -- 6.3.3.8. Numerical results -- 6.3.4. A limit to autonomous approaches: the speed constraint -- 6.4. Iterative approaches -- 6.5. Global approaches -- 6.6. A global approach using evolutionary computation -- 6.6.1. Maneuver modeling -- 6.6.2. Uncertainty modeling -- 6.6.3. Real-time management -- 6.6.4. Evolutionary algorithm implementation -- 6.6.4.1. General description -- 6.6.4.2. The horizon effect -- 6.6.4.3. The fitness function -- 6.6.4.4. Use of partial separability -- 6.6.4.5. The adapted crossover operator -- 6.6.4.6. Theoretical study of a simple example -- 6.6.4.7. Probability of improvement -- 6.6.4.8. Application to conflict resolution -- 6.6.5. Alternative modeling -- 6.6.6. One-day traffic statistics -- 6.6.7. Introducing automation in the existing system -- 6.7. A global approach using ant colony optimization -- 6.7.1. Problem modeling -- 6.7.2. Algorithm description -- 6.7.3. Algorithm improvement: constraint relaxation -- 6.7.4. Results -- 6.7.5. Conclusion and further work -- 6.8. A new framework for comparing approaches -- 6.8.1. Introduction -- 6.8.2. Trajectory prediction model -- 6.8.2.1. Maneuvers -- 6.8.2.2. Decision variables -- 6.8.2.3. Cost -- 6.8.2.4. Handling uncertainties -- 6.8.3. Conflict detection -- 6.8.4. Benchmark generation -- 6.8.5. Conflict resolution -- 6.8.5.1. Evolutionary algorithm -- 6.8.5.1.1. Principles -- 6.8.5.1.2. Fitness function -- 6.8.5.1.3. Adapted crossover and mutation -- 6.8.5.2. Constraint programming -- 6.8.5.2.1. CSP model.

6.8.5.2.2. Solution search -- 6.8.5.3. Optimization -- 6.8.5.4. Results -- 6.8.5.4.1. Benchmark -- 6.8.5.4.2. Conflict resolution -- 6.8.5.4.3. Computing times -- 6.8.5.4.4. Cost of solutions -- 6.8.5.5. Conclusion and further work -- 6.9. Conclusion -- Conclusion -- Bibliography -- Index.