Identification of missing LSBs which can be added in the road network to minimize the travel time

ABSTARCT :

A wide range of relatively short-term disruptive events such as partial flooding, visibility reductions, traction hazards due to weather, and pavement deterioration occur on transportation networks on a daily basis. Despite being relatively minor when compared to catastrophes, these events still have profound impacts on traffic flow. To date there has been very little distinction drawn between different types of network-disruption studies and how the methodological approaches used in those studies differ depending on the specific research objectives and on the disruption scenarios being modeled. In this paper, we advance a methodological approach that employs different link-based capacity-disruption values for identifying and ranking the most critical links and quantifying network robustness in a transportation network. We demonstrate how an ideal capacity-disruption range can be objectively determined for a particular network and introduce a scalable system-wide performance measure, called the Network Trip Robustness (NTR) that can be used to directly compare networks of different sizes, topologies, and connectivity levels. Our approach yields results that are independent of the degree of connectivity and can be used to evaluate robustness on networks with isolating links. We show that system-wide travel-times and the rank-ordering of the most critical links in a network can vary dramatically based on both the capacity-disruption level and on the overall connectivity of the network. We further show that the relationships between network robustness, the capacity-disruption level used for modeling, and network connectivity are non-linear and not necessarily intuitive. We discuss our findings with respect to Braess’ Paradox.

EXISTING SYSTEM :

? The SSNN is a non-linear model with many degrees of freedom there exist (in principle) infinite parameter settings that would fit a particular travel time prediction problem well. ? There exists a wide and detailed body of theoretical and empirical knowledge on the propagation and operation of vehicular traffic. ? We argue that if the number of inputs to a problem is not too large, and either sound theoretical or statistical methods exist to determine the appropriate inputs and input time lags, a feed-forward ANN approach is certainly suitable for dynamic problems.

DISADVANTAGE :

? Network-disruption analysis is a methodological approach that has been successfully applied to transportation maintenance and planning problems to identify and rank the most critical links in a network and to evaluate the robustness of the network as a whole. ? The most obvious problem resulting from completely removing a network links to model disruptions is the creation of isolated sub-networks, which are no longer accessible when the single link connecting them to the network has been removed. ? A shortcoming in the original formulation of the NRI was the inability to address the problem of isolating links in a methodologically sound manner.

PROPOSED SYSTEM :

• Several approximations are proposed that correct for the overestimation of mean speed (and thus underestimation of travel time) through arithmetic time averaging. • We propose robust and easy-to-implement procedures that account for the missing data and allow accurate predictions even at high degrees of input failure. • The central component in the proposed framework is inevitably the data driven short term freeway travel time prediction model, which predicts travel times on the freeway route of interest based on current and / or near-past traffic conditions.

ADVANTAGE :

? It is equally important to have a sound methodological foundation for the assumptions used in the model and to define exactly what performance measures are used and how they are derived. ? Within the domain of network-disruption literature, different network performance measures are frequently used. ? The modified NRI and NTR have advantages over, and increased policy-relevance when compared to well-known and widely used performance measures such as the V/C because they account for the relative value of each link to the entire network as opposed to providing only a localized measure of the value of each link.