1. Decision support tool for operational planning and control in Urban railway traffic
The objective of this project is to develop a decision support tool to help metro developers plan their train schedules, based on demands on various routes, crew schedule and rolling stock management considering effective balancing multi-modal traffic. Given a network alignment, a set of OD pairs and the train operational parameters such as running speed and headways, it should be possible to schedule a few number of services through the network, to generate timetables of each train and to generate station / track occupancy tables. It is also possible to estimate resource demands, when such decision data are available. In addition, timetables can be generated using the tool. We concede to deploy intelligent agents for the DSS, in the current stage of development. Multiple functionalities can be incorporated in the tool, as the implementation progresses. Several known algorithms, such as Shortest path, least span time, greedy algorithms, dynamic programming shall be used in sequencing the operations. Bespoke heuristic procedures shall be developed and deployed, as the project progresses.
A working prototype has been developed and testing is on. Email for more info.
2. Contingency management in logistic operations
Operational contingencies are very common phenomena in several domains. Restoration of operations has been the immediate goal of contingency handling and management. A common way of solving problems of this kind is to perform scenario analysis, also referred to as simulation. Possible future occurrences are sampled as decision variables and the corresponding problem is solved using currently known parametric values. Early models of decision-making under uncertainty relied on statistical decision theory, which extended its scope from a simple expected value to minimax, regret and mean-risk models. Recent modelling paradigms that have emerged as the basis for optimization under uncertainty are recourse, chance-constrained and robust optimization. The concept of robustness in general is to capture flexibility in sequential decision processes, or it could be to represent volatility in contingency events. I am interested in stochastic modelling of contingencies and if possible explore evaluation of stochastic programming models with and without preprocessing. Existing researches attempt to unify stochastic programming models with robust optimization as an overall goal, rather than as an independent model that competes with other modelling paradigms. Various factors that may be addressed in stochastic programming models are sources of uncertainty, time scale of decisions, costs of outcome, decision scope and complexity metrics.
3. Movement planning and dispatching in multi-track territories
Given a dispatching territory, the task of a dispatcher is to assign track authorities to trains for specific times to maximize the overall system efficiency while abiding by a number of safety and desirability constraints. This involves deciding which train takes the siding when two trains travelling in opposite directions meet each other, or when a faster train is to overtake the leading slower train. The dispatcher must also decide on the exit and hold times at all arc segments of the territory. Typically, in a movement planning problem, the territory consists of few miles of single track and double track segments with sidings on each. Maintenance of Way (MOW) window is well-defined on these segments, at specific mileposts at specific times and durations, during which train movements are forbidden. Trains are of different physical and operational features which further impose some strict and preferred planning restrictions. The objective is to develop a deadlock-free, feasible movement plan that minimizes train delays, maximizes schedule adherence, minimizes terminal want-time delays and promotes movements on preferred tracks.
4. Dynamic pricing of transportation services under competition induced uncertainty
The recent economic downturn has brought in severe volatility in pricing - revenue management and portfolio management decisions and has brought down heavily on consumption and consequently on transportation fares. Particularly airlines operate in a highly competitive space. Several models exist to sell an otherwise unused seat or a cargo space. Some of the quantitative determinants of capacity pricing in transportation include market value, replacement costs and historical costs. Some key cost components are return of capital invested over the life of assets, operating and maintenance costs. This research proposes to study the uncertainty in tactical pricing decisions of such capacity due to depreciation in competition pricing. A few specific characteristics of these assets are (i) they are typically oligopoly; (ii) pricing is highly sensitive to competition pricing; (iii) quality depends on service levels and product features; and (iv) service levels and product features are varied and highly complex.
5. Yard capacity optimization
Optimizing operational plans of a classification yard is very important for freight management as it helps fully utilize the limited resources of its rail network. In Indian freight, scientific approach to yard capacity optimization is not in practice, though much effort is spent in US / EU for automated yard capacity optimization. Typical yard planning performance measures are the number of inbound/outbound trains received/assembled, the number of blocks made, the number of cars handled, or the expected time in system per railcar and a common objective is to minimizing the total waiting time of railcars at the yard. To improve the overall yard throughput, an optimized yard operation plan is highly desirable. However, building a classification yard operation plan is challenging as it covers many interrelated operations and decisions. Given arrival times of the inbound trains, this problem is to find the humping schedule and the departure times of the outbound trains subject to the different operational constraints in the yard. The goal is to minimize the total waiting time of railcars in the yard and maximize the total number of railcar processed during a certain period.
6. Rake link management
Rake link management is about rolling stock optimization. Particularly in IR zones, the number of rakes used in a train service is often split and relinked with other routes. Ideally such plans are cyclic, operated over a week’s plan. The objective of efficient rake linking is to minimize the number of rakes deployed over all the planned routes and also to minimize the dry runs. The problem can be augmented to multi-objective optimization with crew planning can be included along with route planning.
A working prototype has been developed and testing is on. Email for more info.
