
Capturing correlation with a mixed recursive logit model for activity-travel scheduling
Transportation Research Part C, 93:73-291, 2018
-
Highlights:
- Combining recursive logit with mixed logit framework for activity-based modeling.
- Model estimated on real size application within reasonable time.
- Confirmed improved out-of-sample fit.

The load planning problem for double-stack intermodal trains
European Journal of Operational Research, 267(1):107-119, 2018
-
Highlights:
- We deal with a multitude of railcars and container types through loading patterns.
- We solve large instance in reasonable time using a commercial solver.
- Numerical results are for the North American market.

On the similarities between random regret minimization and mother logit: the case of recursive route choice models
Journal of Choice Modeling, 23C:21-33, 2017
-
Highlights:
- Link-based route choice models (no path choice set generation).
- We formulate random regret-based recursive logit models.
- Two new specifications of the random regret functions.
- New mother logit models that are equivalent to the regret-based models.
- Estimation and cross-validation results using real data.

A dynamic programming approach for quickly estimating large network-based MEV models
Transportation Research Part B 98(1):179-197, 2017
-
Highlights:
- A new approach for quickly estimating MEV models with large choice sets.
- MEV models are formulated as dynamic discrete choice models.
- The dynamic models generalize the network MEV and are consistent with MEV theory.
- Estimation using the concept of network flows and the nested fixed point algorithm.
- Numerical results based on simulated and real data.

Bike route choice modeling using GPS data without choice sets of paths
Transportation Research Part C, 75(1):183-196, 2017
-
Highlights:
- Model estimated and validated using bike GPS data on network of 40,000 links.
- Quick computation of link flows without choice set generation.
- Easy to compute accessibility measure with desirable properties.

An empirical study on aggregation of alternatives and its influence on prediction in car type choice models
Transportation (2017). https://doi.org/10.1007/s11116-017-9828-5

A decomposition method for estimating recursive logit based route choice models
Euro J Transp Logist (2016). doi:10.1007/s13676-016-0102-3

A misspecification test for logit-based route choice models
Economics of Transportation 4(4):215-226, 2015.
-
Highlights:
- We show how White's information matrix test can be applied to route choice models.
- Results for path-based and link-based (recursive logit) models.
- Comparison of information matrix and IIA and McFadden–Train Lagrange multiplier tests.
- Monte Carlo experiment results assess the size and power of the tests.
- Path-based and link-based models are strongly rejected when applied to real data.

A nested recursive logit model for route choice analysis
Transportation Research Part B 75(1):100-112, 2015
-
Highlights:
- The IIA property is relaxed and random terms are correlated.
- Link-based model: no choice sets of paths are required.
- Efficient estimation and prediction.
- Estimation and cross-validation results using real data.

A link based network route choice model with unrestricted choice set
Transportation Research Part B 56(1):70-80, 2013
-
Highlights:
- A model for the choice of route in a network.
- Link size attribute: a deterministic correction for correlation.
- No choice set generation needed.
- A dynamic specification of sequential link choices is equivalent to a static logit model.
ARTICLES SOUMIS
Hewitt, M., Frejinger, E. Data-driven Optimization Model Customization, soumis à Management Science.
Zimmermann, M., Frejinger, E., Marcotte, P. A strategic Markovian traffic equilibrium model for capacitated networks, soumis à Transportation Science.