The OMDS/BALOR Talk Series invites you to two seminar talks in January 2025:
- Speaker: Dr. Ruben D’Haen (Hasselt University)
- Title: Heuristic algorithms for the online integrated optimisation of order picking and vehicle routing decisions
- Time: Wednesday, January 8, 2025, 15:00 - 16:00
- Location: Seminar Room 5, Kolingasse
- Abstract: How can customer orders be handled in a quick but efficient manner? Due to fierce competition, responding to new customer orders as soon as possible has become essential. Still, optimising operations is indispensable to achieve high service levels at reasonable cost. During my PhD, we studied how the order handling process can be optimised, with a focus on the order picking and vehicle routing problems. To reduce the delivery delay as much as possible, while still attaining efficient operations, new optimisation algorithms are required. A first area of improvement is shifting from static to online optimisation. A second improvement is the use of integrated optimisation, where interrelated processes are optimised at the same time to consider their interaction.
- First, we developed a metaheuristic algorithm to optimise the order picking operations in a B2B context. The interaction between several order picking subproblems, i.e., order batching, picker routing and batch scheduling is considered. In our experiments, we show the benefits of anticipating on future order arrivals, by including dummy orders in the schedule.
- In a second step, we performed a real-life study for a spare-parts company in Belgium. We used real order data and the actual warehouse layout in our algorithm to offer advice to the company. The results show that large operational improvements can be obtained by using our integrated optimisation approach compared to the current operational policy.
- Third, for a B2C e-commerce setting, we show the advantage of optimising order picking and vehicle routing decisions in an integrated manner. Four metaheuristic algorithms, with increasing levels of integration, are developed and tested. The results, once again, highlight the benefits of using integrated algorithms, though at the cost of increased complexity. Finally, the best algorithm is tested in different settings to obtain insights into the challenges of same-day delivery operations, and the impact customers have on the operations.
- Speaker: Prof. Tal Raviv (Tel Aviv University)
- Title: The dynamic bus scheduling problem: Flexibility in a fixed-line transit system
- Time: Wednesday, January 15, 2025, 15:30 - 16:30
- Location: Seminar Room 5, Oskar-Morgenstern-Platz 1
- Abstract:
- Providing adequate public transit service in sparsely populated regions is typically very costly. Low demand implies low vehicle utilization, resulting in a high cost per passenger trip. In many countries, the government subsidizes the service at a high cost to the taxpayers.
- In recent years, many attempts have been made to economize the public transit service by replacing fixed-line buses with mobility-on-demand services, where passengers prebook their trips and vehicles are launched to serve them accordingly. However, this approach leads to even greater spatial and temporal fragmentation of the demand, which is becoming even more challenging to meet using limited resources. Indeed, it is known that legacy fixed bus lines induce demand consolidation as the passengers adapt their daily activities to the fixed schedules of the service and, in the long run, even adapt the location of their activities (workplace, education institute, shopping, etc.) to the public transit routes. Yet, in rural areas, the buses that serve these lines frequently travel empty or nearly empty.
- In this study, we introduce and evaluate a rural transportation service based on fixed lines with fixed schedules where Passengers must book their trips in advance. The lines are served using a heterogeneous fleet, with vehicles of various capacities assigned to the routes dynamically in real-time, and unnecessary trips are canceled. Such a service leverages the benefits of mobility-on-demand systems by matching supply to demand while retaining the advantages of a fixed bus line network. It facilitates the consolidation of passenger journeys through predetermined routes and schedules.
- We verified the potential passenger acceptance of the proposed service in rural communities in Israel through a survey study. Next, we checked the economic viability of the proposed public transit model by formulating the resulting dynamic bus scheduling and stochastic fleet sizing problems as mixed integer linear problems. The dynamic bus scheduling problem is solved with a rolling horizon each time a vehicle dispatching decision is due, based on the currently known demand plus a future forecast updated in real-time.
- Our models were applied to the historical demand faced by a cluster of lines around Karmiel (a town in northern Israel), which serves, on average, about 3000 passenger trips a day. The resulting fleet procurement and operation costs were compared with an idealized version of the current practice with a static vehicle schedule.
- Our experiment shows that switching from static to dynamic bus schedules and redesigning the fleet size and composition accordingly can save about 25% of the costs, which in our use case sum up to about five million shekels a year (approximately 1.25 million Euros). Assuming the public transit budget is fixed, the remaining resources may be used to improve the frequencies of the existing lines, which in turn is likely to increase ridership. Furthermore, under dynamic bus scheduling, the optimal fleet consists of significantly smaller vehicles, reducing emissions and other negative externalities.
- The presentation is based on a joint work with Dr. Ilit Oppenheim from the Shlomo Shemltzer Institute of Smart Transportation at Tel Aviv University and with the entrepreneur Mr. Israel Rom from Operati LTD.
About the Speaker: Associate Professor Tal Raviv is a member of the Industrial Engineering department at the Iby and Aladar Fleischman Faculty of Engineering, Tel Aviv University, Israel. He holds a BA in Economics (1993), and MBA (1997) from Tel Aviv University. He finished his Ph.D. in Operations Research at the Faculty of Industrial Engineering and Management in Technion - Israel Institute of Technology, Haifa (2003). He spent two years (2004-2006) as a postdoctoral fellow at the Sauder School of Business at the University of British Columbia, Vancouver, Canada. Professor Raviv publishes regularly in operational research and transportation science journals, is an advisor for start-up companies, and is an editorial board member in Transportation Science and Transportation Research Parts B. His primary research interest is transportation and logistics, focusing on intelligent and sustainable transportation. He studies shared mobility systems, small parcel delivery logistics, warehousing and intra-logistics, public transit planning, and parking policy. Raviv is co-heading the transportation and logistics lab at Tel Aviv University and has been the head of the Shlomo Shmeltzer Institute for Smart Transportation at Tel Aviv University since it was founded in 2018.
