Adaptive least-expected time paths in stochastic, time-varying transportation and data networks

Abstract

In congested transportation and data networks, travel (or transmission) times are time‐varying quantities that are at best known a priori with uncertainty. In such stochastic, time‐varying (or STV) networks, one can choose to use the a priori least‐expected time (LET) path or one can make improved routing decisions en route as traversal times on traveled arcs are experienced and arrival times at intermediate locations are revealed. In this context, for a given origin‐destination pair at a specific departure time, a single path may not provide an adequate solution, because the optimal path depends on intermediate information concerning experienced traversal times on traveled arcs. Thus, a set of strategies, referred to as hyperpaths, are generated to provide directions to the destination node conditioned upon arrival times at intermediate locations. In this paper, an efficient label‐setting‐based algorithm is presented for determining the adaptive LET hyperpaths in STV networks. Such a procedure is useful in making critical routing decisions in Intelligent Transportation Systems (ITS) and data communication networks. A side‐by‐side comparison of this procedure with a label‐correcting‐based algorithm for solving the same problem is made. Results of extensive computational tests to assess and compare the performance of both algorithms, as well as to investigate the characteristics of the resulting hyperpaths, are presented. An illustrative example of both procedures is provided. © 2001 John Wiley & Sons, Inc.

ICNavS: a tool for reliable dynamic route guidance

“…While this is not the same as seeking the most reliable route, the conjecture is that this is a useful heuristic in a context of limited information about travel time distributions and limited computing power. The task of finding the most reliable route would require knowledge of link travel time distributions and their correlations, which must then be convoluted to produce route travel time distributions, an extremely computationally demanding task [22][23][24][25].…”

Eco-reliable path finding in time-variant and stochastic networks

“…To our knowledge, the existing researches do not specifically consider both the dynamics and randomness of the travel time and emissions in the path finding problems. In the literature, for the path-finding process in time-variant and stochastic networks, many optimal strategies related to travel time have been proposed with various evaluation indices, such as the least expected travel time (Miller-Hooks [24]; Wang et al [35]; Yang and Zhou [38], Yang et al [41]), the probability of on-time arrival (Fan et al [11]; Nie and Wu [27]; Samaranayake et al [32]), the minimum possible travel time (Nielsen [26]), etc. Among them, the least expected travel time method has been widely applied in route-finding studies.…”

“…Among them, the least expected travel time method has been widely applied in route-finding studies. For example, Miller-Hooks [24] used a priori least-expected time path and a set of strategies to improve routing decisions. Yang and Zhou [38] adopted a Lagrangian relaxation-based solution algorithm to reformulate and solve the least expected time path model.…”

Chapter 9 Hazardous Materials Transportation

“…They are both time-dependent and stochastic in nature; i.e., they are random variables with probability distribution functions that vary with time. Therefore, the local routing and scheduling problem is best modeled as a path selection problem in a stochastic time-varying network (see, for example, Hall, 1986;Fu and Rilett, 1998;Miller-Hooks and Mahmassani, 1998;Miller-Hooks, 2001). …”