• Proceedings of the KOR-KST Conference
• /
• 2000.02a
• /
• pp.25-47
• /
• 2000

In route guidance systems fastest-path routing has typically been adopted because of its simplicity. However, empirical studies on route choice behavior have shown that drivers use numerous criteria in choosing a route. The objective of this study is to develop computationally efficient algorithms for identifying a manageable subset of the nondominated (i.e. Pareto optimal) paths for real-time vehicle routing which reflect the drivers' preferences and route choice behaviors. We propose two pruning algorithms that reduce the search area based on a context-dependent linear utility function and thus reduce the computation time. The basic notion of the proposed approach is that ⅰ) enumerating all nondominated paths is computationally too expensive, ⅱ) obtaining a stable mathematical representation of the drivers' utility function is theoretically difficult and impractical, and ⅲ) obtaining optimal path given a nonlinear utility function is a NP-hard problem. Consequently, a heuristic two-stage strategy which identifies multiple routes and then select the near-optimal path may be effective and practical. As the first stage, we utilize the relaxation based pruning technique based on an entropy model to recognize and discard most of the nondominated paths that do not reflect the drivers' preference and/or the context-dependency of the preference. In addition, to make sure that paths identified are dissimilar in terms of links used, the number of shared links between routes is limited. We test the proposed algorithms in a large real-life traffic network and show that the algorithms reduce CPU time significantly compared with conventional multi-criteria shortest path algorithms while the attributes of the routes identified reflect drivers' preferences and generic route choice behaviors well.

• Journal of Korean Society of Transportation
• /
• v.26 no.1
• /
• pp.145-153
• /
• 2008

Existing shortest-path algorithms mainly consider a single attribute. But traveler actually chooses a route considering not single attribute but various attributes which are synthesized travel time, route length, personal preference, etc. Therefore, to search the optimal path, these attributes are considered synthetically. In this study route searching algorithm which selects the maximum utility route using discrete choice model has developed in order to consider various attributes. Six elements which affect route choice are chosen for the route choice model and parameters of the models are estimated using survey data. A multinomial logit models are developed to design the function of route choice model. As a result, the model which has route length, delay time, the number of turning as parameter is selected based on the significance test. We use existing shortest path algorithm, which can reflect urban transportation network such as u-turn or p-turn, and apply it to the real network.

• Journal of Korean Society of Transportation
• /
• v.5 no.1
• /
• pp.3-23
• /
• 1987

Traffic signal control problems in urban street networks are formulated in two ways. In the formulations network flows are assumed to satisfy the user route choice criterion. the first formulation which is called implicit substiuation incorporates user route behavior implicitly in the objective function by recognizing the dependence of the link flows on the signal variables. On the other hands, the second one which is called 'penalty formulation' consists in expressing the route choice conditions in the form of a single nonlinear constraint. Approximate solution algorithm for each of the formulations was investigated in detail and computer codes were written to examine key aspects of each algorithm. A test was done on a network which is small in size but sufficiently complex in representing real-world traffic conditions and the test result shows that both algorithms produce converged solutions. It is recommended, however, that further studies should be done in order to compare the performance of each algorithm more in depth.

• Journal of the Korean Society for Railway
• /
• v.17 no.6
• /
• pp.433-442
• /
• 2014

This paper evaluates the performance of transit assignment algorithms for urban rail networks. The accuracy of the algorithms is essential not just for travel forecasting but also for the area of applications such as the assessment of road vulnerability and the fare adjustments between train operating companies. Nonetheless, the suitability and caveats for the series of computational steps have not yet been much discussed. This study thus considers the characteristics that are appropriate for investigating Seoul rail travelers using three representative transit assignment algorithms: the optimal strategy algorithm, route choice algorithms, and the Dial's algorithm. Both the theoretical foundation and the empirical performance are examined. The results demonstrate that the Dial's algorithm is superior in terms of the theoretical soundness and the computational efficiency.

• Journal of Korean Society of Transportation
• /
• v.22 no.4 s.75
• /
• pp.85-96
• /
• 2004

Given a Path represented by a sequence of link numbers in a transportation network, the reasonable path is defined as a path that any link is appeared multiple times in it. Application of the link labelmethod(LLM) to the shortest path algorithms(SPA) enables to model the reasonable path choice behavior in urban networks. This study aims at expanding the LLM to a Kth shortest path algorithms(KPSA), which adopts the node label setting method. The small-scaled network test demonstrated that the proposed algorithm works correctly and the revised Sioux fall network test showed that the path choice behaviors are reasonably reflected. In the large-scaled network based on the South Korea peninsula, drivers' route diversion perceptions are included as cost terms in total cost. The algorithm may be applied as an alternative route information tools for the deployment of ATIS.

• Journal of Korean Society of Transportation
• /
• v.11 no.1
• /
• pp.67-86
• /
• 1993

A formulation of dynamic traffic assignment between multiple origins and single destination was first introduced in 1987 by Merchant and Nemhauser, and then expanded for multiple destination in the late 1980's (Carey, 1987). Based on behavioral choice theory which provides proper demand elasticities with respect to changes in policy variables, traffic phenomena can be analysed more realistically, especially in peak periods. However, algorithms for these models are not well developed so far(working with only small toy network) and solutions of these models are not unique. In this paper, a new model is developed which keeps the simplicity of static models, but provides the sensitivity of dynamic models with changes of O-D flows over time. It can be viewed as a joint departure time and route choice model, in the given time periods(6-7, 7-8, 8-9 and 9-10 am). Standard multinomial logit model has been used for simulating the choice behavior of destination, mode, route and departure time within a framework of the incremental network assignment model. The model developed is workable in a PC 386 with 175 traffic zones and 3581 links of Seoul and tested for evaluating the exclusive use of Namsan tunnel for HOV and the left-turn prohibition. Model's performance results and their statistical significance are also presented.

• Proceedings of the KOR-KST Conference
• /
• 2000.02a
• /
• pp.89-115
• /
• 2000

Traffic signal setting policies and traffic assignment procedures are mutually dependent. The combined signal control and traffic assignment problem deals with this interaction. With the total travel time minimization objective, gradient based local search methods are implemented. Deterministic user equilibrium is the selected user route choice rule, Webster's delay curve is the link performance function, and green time per cycle ratios are decision variables. Three implemented solution codes resulting in six variations include intersections operating under multiphase operation with overlapping traffic movements. For reference, the iterative approach is also coded and all codes are tested in four example networks at five demand levels. The results show the numerical gradient estimation procedure performs best although the simplified local searches show reducing the large network computational burden. Demand level as well as network size affects the relative performance of the local and iterative approaches. As demand level becomes higher, (1) in the small network, the local search tends to outperform the iterative search and (2) in the large network, vice versa.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.7 no.1
• /
• pp.10-21
• /
• 2008

User travel behavior is based on the existence of complete traffic information in deterministic user optimal principle by Wardrop(1952). According to deterministic user optimal principle, users choose the optimal route from origin to destination and they change their routes arbitrarily in order to minimize travel cost. In this principle, users only consider travel time as a factor to take their routes. However, user behavior is not determined by only travel time in actuality. Namely, the models that reflect only travel time as a route choice factor could give irrational travel behavior results. Therefore, the model is necessary that considers various factors including travel time, transportation networks structure and traffic information. In this research, more realistic deterministic optimal traffic assignment model is proposed in the way of route recognizance behavior. This model assumes that when users decide their routes, they consider many factors such as travel time, road condition and traffic information. In addition, route recognizance attributes is reflected in this suggested model by forward searching method and backward searching method with numerical formulas and algorithms.

• Journal of Korean Society of Transportation
• /
• v.22 no.6
• /
• pp.121-131
• /
• 2004

The K-shortest path algorithms are largely classified into two groups: oneis for finding loopless path (simple path), another loop paths. In terms of cimputational complexities, in general the loop-paths-finding ones are considered more efficient and easier to be handled than the loopless-paths-finding. The entire path deletion methods have been known as the best efficient algorithms among the proposed K-shortest path algorithms. These algorithms exploit the K-th network transformation to prevent the same path, which was already selected as the (K-1)th path, from being redetected. Nevertheless, these algorithms have a critical limitationto be applied in the practical traffic networks because the loops, in which the same modes and links can be unlimitedly repeated, are not preventable. This research develops a way to be able to selectively control loop-paths by applying link-label. This research takes an advantage of the link-based shortest path algorithms that since the algorithms can take care of two links simultaneouslyin the searching process, the generation of loops can be controlled in the concatenation process of the searched link and the preceded link. In concatenation of two links, since the precede link can be treated a sub-shortest to this link from the origination, whether both the node and the link of the searched link were already existed or not can be evaluated. Terefore, both the node-loopless path, in which the same node is not appeared, and the link-loopless, in which the same link is not appeared, can be separately controlled. Especially, the concept of the link-loopless path is expended to take into consideration reasonable route choice behaviors such as U-Turn, P-Turn, and Turn-Penalty, which are frequently witnessed in urban traffic network with intersections. The applicability of the proposed method is verified through case studies.

• The Transactions of the Korean Institute of Electrical Engineers D
• /
• v.49 no.7
• /
• pp.378-388
• /
• 2000

In this paper, we propose Fuzzy Polynomial Neural Networks(FPNN) based on Polynomial Neural Networks(PNN) and Fuzzy Neural Networks(FNN) for model identification of complex and nonlinear systems. The proposed FPNN is generated from the mutually combined structure of both FNN and PNN. The one and the other are considered as the premise part and consequence part of FPNN structure respectively. As the consequence part of FPNN, PNN is based on Group Method of Data Handling(GMDH) method and its structure is similar to Neural Networks. But the structure of PNN is not fixed like in conventional Neural Networks and self-organizing networks that can be generated. FPNN is available effectively for multi-input variables and high-order polynomial according to the combination of FNN with PNN. Accordingly it is possible to consider the nonlinearity characteristics of process and to get better output performance with superb predictive ability. As the premise part of FPNN, FNN uses both the simplified fuzzy inference as fuzzy inference method and error back-propagation algorithm as learning rule. The parameters such as parameters of membership functions, learning rates and momentum coefficients are adjusted using genetic algorithms. And we use two kinds of FNN structure according to the division method of fuzzy space of input variables. One is basic FNN structure and uses fuzzy input space divided by each separated input variable, the other is modified FNN structure and uses fuzzy input space divided by mutually combined input variables. In order to evaluate the performance of proposed models, we use the nonlinear function and traffic route choice process. The results show that the proposed FPNN can produce the model with higher accuracy and more robustness than any other method presented previously. And also performance index related to the approximation and prediction capabilities of model is evaluated and discussed.