A stochastic multi‐period investment selection model to optimize strategic railway capacity planning

Abstract

SUMMARYNorth American Freight Railroads are approaching the limits of practical capacity because of substantial future demand. In this research, we develop a Stochastic Multi-period Investment Selection Model (S-MISM) to assist railroads best allocate their capital investments in the long-term strategic capacity planning process. The novel optimization framework uses stochastic programming and Benders decomposition and provides a means to cope with unfulfilled demand and demand uncertainty in a long-term multi-period investment selection problem. S-MISM can determine which portions of a rail network need to be upgraded with what kind of expansion options at each defined period in the planning horizon. Experimental results show that the inclusion of demand uncertainty results in a better and more robust capacity plan. Using this decision support tool will help railroads maximize their return from capacity expansion projects and minimize the risk in strategic capacity planning subject to demand uncertainty.

Multistage stochastic programming model and solution algorithm for the capacity expansion of railway network

“…The structure of the constraints is illustrated in Fig 1. Aggregate level decision Detailed level decision Aggregate level decision Detailed level decision The L-shaped method is often applied to this type of problem. Indeed, in Lai and Shih (2013), it was applied to the railway expansion problem. It is true that they showed its effectiveness by numerical experiment, but they did not provide the theoretical backgrounds such as a proof of convergence.…”

“…Although the feasibility of the problem is satisfied, the demand may not be satisfied in the conventional method by Lai and Shih (2013). To satisfy the demand constraints, the value of Shiina, Takaichi, Li, Morito and Imaizumi, Journal of Advanced Mechanical Design, Systems, and Manufacturing, Vol.12, No.3 (2018) the penalty parameter τ must be set to sufficiently large.…”

“…The nework shown in Fig. 4 is based on Lai and Shih (2013), where the number on the arc denotes the number of arcs in forward direction (that in backward direction is in parentheses). The origin and the destination of OD pair is shown in Table 1.…”

“…In our L-shaped method, a parameter controls the penalty value. In this paper, we show how to set the value of the parameter so as to eventually satisfy the demands, which has not been mentioned clearly in the previous research by Lai and Shih (2013). By adding the optimality cut, which approximates the penalty, to the master problem, the capacity expansion is chosen to reduce the penalty for unsatisfied demand.…”

“…Demand for cargo transport via railway in North America is on the rise. According to Lai and Barkan (2011), the demand is expected to increase by 84% in 2035, compared to 2007. Lai and Shih (2013 dealt with the capacity expansion problem of railways as a stochastic programming problem to maximize the fulfilled demands for railways, and to extend routes within a limited budget.…”