Machine Learning in Airline Crew Pairing to Construct Initial Clusters for Dynamic Constraint Aggregation

The crew pairing problem (CPP) is generally modelled as a set partitioning problem where the flights have to be partitioned in pairings. A pairing is a sequence of flight legs separated by connection time and rest periods that starts and ends at the same base. Because of the extensive list of complex rules and regulations, determining whether a sequence of flights constitutes a feasible pairing can be quite difficult by itself, making CPP one of the hardest of the airline planning problems. In this paper, we first propose to improve the prototype Baseline solver of Desaulniers et al. (2020) by adding dynamic control strategies to obtain an efficient solver for large-scale CPPs: Commercial-GENCOL-DCA. These solvers are designed to aggregate the flights covering constraints to reduce the size of the problem. Then, we use machine learning (ML) to produce clusters of flights having a high probability of being performed consecutively by the same crew. The solver combines several advanced Operations Research techniques to assemble and modify these clusters, when necessary, to produce a good solution. We show, on monthly CPPs with up to 50 000 flights, that Commercial-GENCOL-DCA with clusters produced by ML-based heuristics outperforms Baseline fed by initial clusters that are pairings of a solution obtained by rolling horizon with GENCOL. The reduction of solution cost averages between 6.8% and 8.52%, which is mainly due to the reduction in the cost of global constraints between 69.79% and 78.11%.

[1]  Muhammet Deveci,et al.  Evolutionary algorithms for solving the airline crew pairing problem , 2018, Comput. Ind. Eng..

[2]  Jasper Snoek,et al.  Practical Bayesian Optimization of Machine Learning Algorithms , 2012, NIPS.

[3]  brahim Özkol,et al.  An Improved Genetic Algorithm for Crew Pairing Optimization , 2013 .

[4]  Blaise Hanczar,et al.  Accuracy-Rejection Curves (ARCs) for Comparing Classification Methods with a Reject Option , 2009, MLSB.

[5]  Tzung-Pei Hong,et al.  A Two-Dimensional Genetic Algorithm and Its Application to Aircraft Scheduling Problem , 2015 .

[6]  Zoubin Ghahramani,et al.  Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning , 2015, ICML.

[7]  Martin W. P. Savelsbergh,et al.  Branch-and-Price: Column Generation for Solving Huge Integer Programs , 1998, Oper. Res..

[8]  Edward K. Baker,et al.  Efficient heuristic algorithms for the weighted set covering problem , 1981, Comput. Oper. Res..

[9]  Tara N. Sainath,et al.  Improving deep neural networks for LVCSR using rectified linear units and dropout , 2013, 2013 IEEE International Conference on Acoustics, Speech and Signal Processing.

[10]  Nitish Srivastava,et al.  Dropout: a simple way to prevent neural networks from overfitting , 2014, J. Mach. Learn. Res..

[11]  Ali Azadeh,et al.  A hybrid meta-heuristic algorithm for optimization of crew scheduling , 2013, Appl. Soft Comput..

[12]  Pierre Hansen,et al.  Stabilized column generation , 1998, Discret. Math..

[13]  Partha Kumar Pandit,et al.  Business model of aircraft fleet planning using ANN , 2018 .

[14]  Cynthia Barnhart,et al.  Improving Crew Scheduling by Incorporating Key Maintenance Routing Decisions , 2003, Oper. Res..

[15]  Shengli Qiu,et al.  Airline crew pairing optimization problems and capacitated vehicle routing problems , 2012 .

[16]  Sergey Ioffe,et al.  Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift , 2015, ICML.

[17]  Sander Bohte,et al.  Conditional Time Series Forecasting with Convolutional Neural Networks , 2017, 1703.04691.

[18]  John R. Birge,et al.  A Stochastic Programming Approach to the Airline Crew Scheduling Problem , 2006, Transp. Sci..

[19]  Guy Desaulniers,et al.  Multi-phase dynamic constraint aggregation for set partitioning type problems , 2010, Math. Program..

[20]  Navdeep Jaitly,et al.  Pointer Networks , 2015, NIPS.

[21]  François Soumis,et al.  Improved Primal Simplex: A More General Theoretical Framework and an Extended Experimental Analysis , 2015, INFORMS J. Comput..

[22]  Silvano Martello,et al.  Meta-Heuristics: Advances and Trends in Local Search Paradigms for Optimization , 2012 .

[23]  L. Bianco,et al.  A heursitic procedure for the crew rostering problem , 1992 .

[24]  Jacques Desrosiers,et al.  Chapter 2 Time constrained routing and scheduling , 1995 .

[25]  Jacques Desrosiers,et al.  Accelerating Strategies in Column Generation Methods for Vehicle Routing and Crew Scheduling Problems , 2002 .

[26]  Guy Desaulniers,et al.  Dynamic Aggregation of Set-Partitioning Constraints in Column Generation , 2003, Oper. Res..

[27]  Guy Desaulniers,et al.  A new heuristic branching scheme for the crew pairing problem with base constraints , 2016, Comput. Oper. Res..

[28]  François Soumis,et al.  Airline crew scheduling: models, algorithms, and data sets , 2014, EURO J. Transp. Logist..

[29]  François Soumis,et al.  Flight-connection Prediction for Airline Crew Scheduling to Construct Initial Clusters for OR Optimizer , 2020, ArXiv.

[30]  Marco E. Lübbecke,et al.  Learning When to Use a Decomposition , 2017, CPAIOR.

[31]  Franck Dernoncourt,et al.  Optimizing neural network hyperparameters with Gaussian processes for dialog act classification , 2016, 2016 IEEE Spoken Language Technology Workshop (SLT).

[32]  Jonathan F. Bard,et al.  Flexible weekly tour scheduling for postal service workers using a branch and price , 2013, J. Sched..

[33]  Jacques Desrosiers,et al.  A Column Generation Approach for Large-Scale Aircrew Rostering Problems , 1999, Oper. Res..

[34]  Jacques Desrosiers,et al.  Crew Pairing at Air France , 1993 .

[35]  Kevin Leyton-Brown,et al.  Sequential Model-Based Optimization for General Algorithm Configuration , 2011, LION.

[36]  Guy Desaulniers,et al.  Dynamic Constraint Aggregation for Solving Very Large-scale Airline Crew Pairing Problems , 2020, SN Operations Research Forum.

[37]  Nima Hatami,et al.  Classification of time-series images using deep convolutional neural networks , 2017, International Conference on Machine Vision.

[38]  Martin Desrochers,et al.  A Column Generation Approach to the Urban Transit Crew Scheduling Problem , 1987, Transp. Sci..

[39]  Guy Desaulniers,et al.  Aircrew pairings with possible repetitions of the same flight number , 2009, Comput. Oper. Res..

[40]  Jacques Desrosiers,et al.  Routing with time windows by column generation , 1983, Networks.

[41]  Guy Desaulniers,et al.  An Improved Primal Simplex Algorithm for Degenerate Linear Programs , 2007, INFORMS J. Comput..

[42]  Ellis L. Johnson,et al.  A Global Approach to Crew-Pairing Optimization , 1992, IBM Syst. J..

[43]  François Soumis,et al.  Dynamic constraint and variable aggregation in column generation , 2014, Eur. J. Oper. Res..

[44]  Tsuyoshi Murata,et al.  {m , 1934, ACML.

[45]  Roy E. Marsten,et al.  Exact solution of crew scheduling problems using the set partitioning model: Recent successful applications , 1981, Networks.

[46]  Ruslan Sadykov,et al.  Recent results for column generation based diving heuristics , 2016 .

[47]  Andrea Lodi,et al.  Learning to Handle Parameter Perturbations in Combinatorial Optimization: an Application to Facility Location , 2019, EURO J. Transp. Logist..

[48]  François Soumis,et al.  Improved integral simplex using decomposition for the set partitioning problem , 2018, EURO J. Comput. Optim..

[49]  P.-Q Pan A Basis-Deficiency-Allowing Variation of the Simplex Method for Linear Programming , 2003 .