Dynamic Multi-period Freight Consolidation

Rivera, Arturo E. Pérez; Mes, Martijn

doi:10.1007/978-3-319-24264-4_26

Cited by 6 publications

(10 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Remind that an instance represents a transportation network with many possible states, as seen in Table 2. We use the algorithm settings recommended in the literature [Pérez Rivera andMes, 2015, Powell, 2007]. Second, we test the resulting ADP policy (of each state, per instance and VFA combination) in a simulation of 500 replications of the time horizon.…”

Section: Methodsmentioning

confidence: 99%

Anticipatory freight selection in intermodal long-haul round-trips

Rivera

Mes

2017

Transportation Research Part E: Logistics and Transportation Re

View full text Add to dashboard Cite

We consider a Logistic Service Provider (LSP) that transports freight periodically in a long-haul round-trip. At the start of a round-trip, the LSP consolidates freights in a long-haul vehicle and delivers them to multiple destinations in a region. Within this region, the LSP picks up freights using the same vehicle and transports them back to the starting location. The same region is visited every period, independent of which freights were consolidated. Consequently, differences in costs between two periods are due to the destinations visited (for delivery and pickup of freights) and the use of an alternative transport mode. Freights have different time-windows and become known gradually over time. The LSP has probabilistic knowledge about the arrival of freights and their characteristics. Using this knowledge, the goal of the LSP is to consolidate freights in a way that minimizes the total costs over time. To achieve this goal, we propose the use of a look-ahead policy, which is computed using an Approximate Dynamic Programming (ADP) algorithm. We test our solution method using information from a Dutch LSP that transports containers daily, by barge, from the East of the country to different terminals in the port of Rotterdam, and back. We show that, under different instances of this real-life information, the use of an ADP policy yields cost reductions up to 25.5% compared to a benchmark policy. Furthermore, we discuss our findings for several network settings and state characteristics, thereby providing key managerial insights about look-ahead policies in intermodal long-haul round-trips.

show abstract

Section: Methodsmentioning

confidence: 99%

Anticipatory freight selection in intermodal long-haul round-trips

Rivera

Mes

2017

Transportation Research Part E: Logistics and Transportation Re

View full text Add to dashboard Cite

show abstract

“…The weight α for ADP 2 is defined as α = max {25/ (25 + n − 1) , 0.05} and the sampling method is the same as the advance sampling procedure introduced in the next paragraph. The ADP algorithm runs for 100 iterations and the NLS parameters used are those recommended by [11].…”

Section: Methodsmentioning

confidence: 99%

“…To incorporate stochasticity in DSND approaches, techniques such as scenario generation [3,7], two-stage stochastic programming [1,8], and approximate dynamic programming (ADP) [4,11] have been used. Although these approaches perform better than their deterministic counterpart, they have limitations when considering synchromodal planning.…”

Section: Literature Reviewmentioning

confidence: 99%

Service and Transfer Selection for Freights in a Synchromodal Network

Rivera¹,

Mes²

2016

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

We study the planning problem of selecting services and transfers in a synchromodal network to transport freights with different characteristics, over a multi-period horizon. The evolution of the network over time is determined by the decisions made, the schedule of the services, and the new freights that arrive each period. Although freights become known gradually over time, the planner has probabilistic knowledge about their arrival. Using this knowledge, the planner balances current and future costs at each period, with the objective of minimizing the total costs over the entire horizon. To model this stochastic and multi-period tradeoff, we propose a Markov Decision Process (MDP) model. To overcome the computational complexity of solving the MDP, we propose an Approximate Dynamic Programming (ADP) approach. Using different problem settings, we show that our look-ahead approach has significant benefits compared to a benchmark heuristic.

show abstract

“…Following Bellman's principle of optimality, the best policy π for the planning horizon can be found solving a set of stochastic recursive equations that consider the current-stage and expected next-stage costs, as seen in (3). We can solve (3) plugging in the transition function (19a) and specifying the probability P(W t+1 = ω), which can be found in Pérez Rivera and Mes (2015).…”

Section: Solutionmentioning

confidence: 99%

“…For further reading on this problem, we refer to Pérez Rivera and Mes (2015). In addition, we refer to Van Heeswijk et al (2015) for a similar ADP approach on a completely different transportation problem.…”

Section: Value Function Approximationmentioning

confidence: 99%