Graduate Student Research Seminar Day ‑ Nov 18, 2026
You are cordially invited to the Graduate Student Research Seminar of the Department of Industrial Engineering.
Date: Wednesday, Nov 18, 2026
Time: 11:00 am - 1:05 pm AST
In Person: Morroy Building ("I"), Room 310, Sexton Campus
Online:ÌýIn-person only
(NOTE:Ìý Students are reminded that they must attend in person if they are planning to put this toward their Seminar course requirements.)
Schedule:
1100-1125 |
Ali Mahmoudi, PhD Candidate Beyond Fleet Replacement: An Equitable Bi-level Game Theoretic Policy-making Framework for Bus Transit Decarbonization with Adaptive Passenger Response |
| 1125-1150 | Pehgah Begheri, PhD Candidate A Traffic-Aware Time-Dependent Vehicle Routing Problem for Middle-Mile Deliveries under Travel Time Uncertainty |
| 1150-1215 | Nadia Aghamohammdtajer, MASc. student Robust Flexible Job-Shop Scheduling with Auxiliary Resources and Reentrant Flows: Application to Semiconductor Photolithography Process |
| 1215-1240 | Nasser Al Musallami, MASc. student, MASc. student Fleet Level Maintenance Scheduling with Flexible Work Periods and Resource Levelling |
| 1240-1305 | Alireza Amjadian, PhD Candidate Repair or Return? Impacts of Pricing, Repair Subsidies, and Refund Strategies |
Abstracts:
Beyond Fleet Replacement: An Equitable Bi-level Game Theoretic Policy-making Framework for Bus Transit Decarbonization with Adaptive Passenger Response
Ali Mahmoudi, PhD Candidate
Decarbonizing urban bus fleets requires not only replacing conventional vehicles with cleaner technologies but also attracting indifferent passengers (i.e., gray passengers) who may otherwise choose private cars or ride-sourcing services. This study develops a bi-level framework, using multi-objective optimization and evolutionary game, to analyze the behaviour of gray passengers in the bus fleet transition problem under competing transportation modes. The leader (public authority) selects a policy portfolio (i.e., green fleet support subsidies, non-green fleet penalties, ride-sourcing levies, congestion charges, and carbon pricing policy). In response, gray passengers adjust their modal choices among bus, private car, and ride-sourcing, while the transit operator adjusts the green share of its fleet through lower-level evolutionary preference dynamics. The framework embraces equity considerations by capturing key determinants of passenger choice in mode selection (i.e., travel time, service frequency, crowding, fare, and ride-comfort attributes such as noise, vibration, thermal comfort, and acceleration harshness). The results show that the optimized policy shifts approximately 59.6% of gray passengers to sustainable urban buses, while 23.9% remain with private cars and 16.5% choose ride-sourcing services. The route-level results reveal substantial heterogeneity, indicating that system-wide policy levies should be complemented by route-sensitive service planning. Sensitivity analyses show that gray-market size primarily affects the magnitude of induced bus demand; congestion charges and ride-sourcing levies should be coordinated to avoid cross-substitution; and green-fleet adoption follows a clear subsidy--cost threshold. The findings suggest that effective transit decarbonization requires an integrated policy design that jointly considers fleet transition, competing mobility services, and adaptive passenger behavior.
Ìý
A Traffic-Aware Time-Dependent Vehicle Routing Problem for Middle-Mile Deliveries under Travel Time Uncertainty
Pehgah Begheri, PhD Candidate
Urban freight distribution operates under traffic conditions shaped by both recurrent congestion and unexpected disruptions, which can reduce delivery reliability and vehicle productivity. These effects are particularly important in middle-mile retail logistics, where delays can disrupt store replenishment schedules and subsequent distribution activities. This study develops a traffic-aware stochastic time dependent vehicle routing problem with time windows (STDVRPTW) for middle-mile retail distribution that jointly represents time-dependent travel times, travel time deviations due to disruptions, customer delivery-time requirements, and endogenous fleet deployment. The problem is formulated as a two-stage stochastic program in which vehicle activation, routing, customer allocation, and planned time-bucket decisions are made before uncertainty is realized, while scenario-dependent service start times and time window deviations are evaluated afterward. An improved exact Benders decomposition is developed and strengthened with problem-specific valid inequalities exploiting the time-bucket and vehicle-capacity structure. The numerical results show that the strengthened decomposition reduces Benders iterations and achieves substantial runtime improvements over standard Benders for medium-to-large instances. The operational analyses further show that congestion pricing can shift deliveries away from peak periods, but its effectiveness depends strongly on delivery-window flexibility. At sufficiently high traffic penalties, activating an additional vehicle can become preferable to further schedule shifting, reducing time-window violations at the expense of higher fleet and total operating costs. These findings highlight the importance of jointly considering traffic-management policies, receiver flexibility, and fleet deployment when planning reliable middle-mile deliveries under uncertain urban travel conditions.
Robust Flexible Job-Shop Scheduling with Auxiliary Resources and Reentrant Flows: Application to Semiconductor Photolithography Process
Nadia Aghamohammdtajer, MASc. student
Semiconductor manufacturing involves long production cycles and tightly coupled processes, with photolithography representing a critical bottleneck. Efficient scheduling is therefore essential to reducing overall lead time. However, existing studies predominantly rely on deterministic formulations that overlook uncertainty in fabrication environments. This paper proposes a robust mixed-integer linear programming (MILP) model for photolithography scheduling under processing time uncertainty. The problem is formulated as a flexible job-shop scheduling problem incorporating reentrant flows, sequence-dependent setup times, reticle constraints, and machine-dependent processing times. A budgeted uncertainty framework is employed to address processing time variability. Three solution approaches are developed and evaluated: (i) an exact Benders decomposition (BD) algorithm with tailored cuts, (ii) a fast heuristic inspired by logic-based Benders decomposition, and (iii) a calibrated genetic algorithm (GA) as a benchmark. Computational experiments on 20 instances demonstrate that BD outperforms the full MILP for medium and large instances. The proposed heuristic achieves near-optimal solutions, with average optimality gaps of 1.25–1.34%, while operating two to four orders of magnitude faster than exact methods. It also generally outperforms GA in both solution quality and computational efficiency. These findings establish the proposed framework as a scalable and practical approach to robust photolithography scheduling.
Fleet Level Maintenance Scheduling with Flexible Work Periods and Resource Levelling
Nasser Al Musallami, MASc. student
Ships alternate between operational missions and short work periods in which maintenance must be carried out. In the context of multi-ship maintenance, overlapping work periods can create large and expensive overloads of shared resources. This research develops a consolidated fleet maintenance scheduling plan to better integrate existing individual ship plans that minimizes penalties incurred by advancement, deferral of maintenance actions, late certifications, and resource overloads. The problem is formulated as a mixed integer linear program in which work periods may shift by up to a set number of weeks and maintenance occurrences may be reassigned to nearby work periods. Experiments on a three-ship fleet with 10,192 maintenance occurrences and about 80,900 labor hours, solved with Gurobi, show that the plan as given carries about 15,090 overload hours. Using the proposed model to Shift work periods by six weeks alone reduces overload to about 1,560 hours. Combining six week shifts with reassignment to any work period brings it down to 2 hours. A sensitivity analysis over shift range and reassignment radius maps the trade off between timing deviations and workload levelling. The results show that coordinating when ships enter their work periods, rather than reducing the maintenance itself, removes most of the fleet overload.
Repair or Return? Impacts of Pricing, Repair Subsidies, and Refund Strategies
Alireza Amjadian, PhD Candidate
This study analyses how consumer-targeted repair subsidies interact with an original equipment manufacturer’s (OEM’s) pricing, refund, repair, and product recovery decisions. An analytical model is developed in which the OEM sets the product price and refund for returned failed products, while consumers decide whether to purchase and, after a post-warranty failure, whether to repair or return. The equilibrium has three post-failure regimes: return only, mixed repair and return, and repair only. The subsidy’s effect depends on the regime. Below the repair-entry threshold, the OEM adjusts its price and refund while repair demand remains zero. In the mixedÌýregime, a higher subsidy shifts failed products from return to repair. How readily this occurs depends on the recovery value of returns, which gives the OEM a stronger incentive to preserve returns, and on repair profitability, which encourages repair. Thus, a subsidy promoting repair may reduce the supply of returned products available for recovery. Higher repair demand also need not raise social welfare when repair costs and recovery value outweigh the after-repair use value. An illustration based on France’s "Bonus Réparations" shows repair and return coexisting. The results guide the joint design of repair subsidies and OEM pricing, refund, repair, and recovery decisions.
Contact Person:
Hamid Afshari, Ph.D., P.Eng.
email: hamid.afshari@dal.caÌý