The Rack­ham Pre­doc­tor­al Fel­low­ship is one of the most pres­ti­gious awards grant­ed by the Rack­ham Grad­u­ate School. Doc­tor­al can­di­dates who expect to grad­u­ate with­in six years since begin­ning their degrees are eli­gi­ble to apply, and the strength and qual­i­ty of their dis­ser­ta­tion abstract, pub­li­ca­tions and pre­sen­ta­tions, and rec­om­men­da­tions are all tak­en into con­sid­er­a­tion when grant­i­ng this award.

For over sev­en­ty years, trans­porta­tion net­work equi­lib­ri­um mod­els have been foun­da­tion­al in trans­porta­tion plan­ning, illus­trat­ing trav­el­er com­pe­ti­tion on con­gest­ed net­works to reach an equi­lib­ri­um state, where no trav­el­er ben­e­fits from chang­ing routes. Orig­i­nat­ing in the 1950s, these mod­els faced lim­i­ta­tions due to scarce trav­el data and sim­pli­fied behav­ioral assump­tions. Today, the emer­gence of vehi­cle-to-every­thing data col­lec­tion tech­nol­o­gy offers an excit­ing oppor­tu­ni­ty to trans­form trans­porta­tion net­work analysis.

Zhichen’s dis­ser­ta­tion seeks to trans­form trans­porta­tion net­work equi­lib­ri­um mod­el­ing by lever­ag­ing mul­ti-source data and machine learn­ing to enhance deci­sion-mak­ing in con­nect­ed trans­porta­tion sys­tems. Specif­i­cal­ly, She estab­lish­es a deep learn­ing-based “end-to-end” net­work equi­lib­ri­um frame­work that mod­els net­work-lev­el trav­el­er inter­ac­tion from data.