Avinash Unnikrishnan

Influence of Demand Uncertainty and Correlations on Traffic Predictions and Decisions

: Decisions to improve a regional transportation network are often based on predictions of future link flows that assume future travel demand is a deterministic matrix. Despite broad awareness of the uncertainties inherent in forecasts, rarely are uncertainties considered explicitly within the methodological framework due at least in part to a lack of knowledge as to how uncertainties affect the optimality of decisions. This article seeks to address this issue by presenting a new method for evaluating future travel demand uncertainty and finding an efficient technique for generating multiple realizations of demand. The proposed method employs Hypersphere Decomposition, Cholesky Decomposition, and user equilibrium traffic assignment. Numerical results suggest that neglecting correlations between the future demands of travel zone pairs can lead to improvement decisions that are less robust and could frequently rank improvements improperly. Of the six sampling techniques employed, Antithetic sampling generated travel demand realizations with the least relative bias and error.

A time of day analysis of crashes involving large trucks in urban areas

Previous studies have looked at different factors that contribute to large truck-involved crashes, however a detailed analysis considering the specific effects of time of day is lacking. Using the Crash Records Information System (CRIS) database in Texas, large truck-involved crashes occurring on urban freeways between 2006 and 2010 were separated into five time periods (i.e., early morning, morning, mid-day, afternoon and evening). A series of log likelihood ratio tests were conducted to validate that five separate random parameters logit models by time of day were warranted. The outcomes of each time of day model show major differences in both the combination of variables included in each model and the magnitude of impact of those variables. These differences show that the different time periods do in fact have different contributing factors to each injury severity further highlighting the importance of examining crashes based on time of day. Traffic flow, light conditions, surface conditions, time of year and percentage of trucks on the road were found as key differences between the time periods.

Design and Management Strategies for Mixed Public Private Transportation Networks: A Meta-Heuristic Approach

This article presents a new multi-objective mathematical programming framework to model interactions between public and private sectors in constructing and maintaining highway networks using the build, operate, and transfer scheme. In this study, private companies are assumed to have a degree of control over highway sections on which they perform maintenance and rehabilitation and capacity expansion activities. Private investors recover the cost of construction by levying tolls. The public agency is assumed to maintain the rest of the network with the aim of minimizing total system generalized cost. The bi-directional impact of roadway utilization on deterioration and deterioration on utilization is modeled in this study. The model accounts for route choice of users and all users are assumed to choose routes that have equal and minimal experienced generalized cost. The nonconvex and discontinuous multi-objective mathematical program is solved using nondominant sorting genetic algorithm-II and the pareto-optimal trade-off surface between the profit of the private company and the total system cost is generated. Computational runs are conducted to demonstrate the suitability and flexibility of the developed framework in modeling various policy decisions such as the presence of noncompete clauses.

Integrated Traffic Simulation-Statistical Analysis Framework for Online Prediction of Freeway Travel Time

This paper introduces a novel approach to the online short-term prediction of point-to-point freeway travel time, combining statistical forecasting techniques with traffic simulation. At every freeway entrance point, a time series analysis model based on traffic detector counts is used to predict traffic demands, whose flow through the freeway segment is simulated by a cell transmission model. This procedure, applied within a rolling-horizon framework, generates online travel time predictions consistent with traffic flow theory. Experimental results obtained from synthetic data strongly suggest that the estimates obtained with this methodology are robust and accurate. For a wide range of congestion conditions and freeway segment lengths, more than half of the predictions errors were found to be smaller than 15%. Moreover, 80% of these errors fell below 40 s when the actual travel times ranged between 3 and 10 min. Further analyses of the model sensitivity to traffic detector coverage revealed that detector separations of approximately 1 mi resulted in the most precise travel time estimates. In addition to its satisfactory performance, the proposed framework is flexible, and it can make use of additional online data and easily incorporate different forecasting and simulation techniques. Therefore, this work provides a powerful tool for online travel time prediction, suitable for a variety of practical implementation conditions and requirements.

Models for Minimizing Backhaul Costs Through Freight Collaboration

The competitive nature of the trucking industry has forced trucking firms to develop innovative solutions to improve their operational efficiency and decrease marginal costs. One way carriers and shippers are accomplishing these tasks is by collaborating on operations through various strategies. Two optimization models are developed to route the carrier of interests backhaul routes and select collaborative shipments to fulfill; one is formulated as an integer program, and the other is formulated as a mixed integer program. Two solution methodologies, a greedy heuristic and tabu search, are used to solve the two problems, and numerical analysis is performed with a real-world freight network. Numerical analysis reveals that the percentage of cost savings for backhaul routes can be as high as 27%.

Online Freight Network Assignment Model with Transportation Disruptions and Recourse

Continuous real-time monitoring of shipment cost and delivery time is increasingly used by companies to reduce transportation costs while avoiding costly delays or disruptions. Recourse refers to the ability of the shipper to update a routing strategy on the basis of information obtained about the state of the network disruptions. An adaptive routing policy is proposed to help shippers and carriers save costs by reacting to information updates. Public transportation agencies can use the formulation to predict the behavior of shippers under disruptions in multimodal transportation networks. A mathematical model is formulated and analyzed; the model accounts for a new type of freight network assignment problem with recourse defined in a dynamic environment and in the presence of probable network disruptions or significant delays. The mathematical model is intended (a) to capture shipper behavior better in the presence of network disruptions and rerouting and (b) to include the heterogeneity in shipper routing behavior as a result of commodity or product, mode, and logistics system characteristics. Results indicate that models that do not account for the likelihood of disruption can severely misrepresent freight system performance and flows over the network, especially in situations in which freight is continuously monitored and disruptions are either significant or frequent.

Robust Optimization Strategy for the Shortest Path Problem under Uncertain Link Travel Cost Distribution

This article showed how numerical experiments conducted on small to large networks compare the robust optimization-based strategy to the classical deterministic shortest path in terms of the uncertainty. A robust optimization approach for the shortest path problem where travel cost is uncertain and exact information on the distribution function is unavailable is developed. The article showed that under such conditions the robust shortest path problem can be formulated as a binary nonlinear integer program, which can then be reformulated as a mixed integer conic quadratic program. This article presented an outer approximation algorithm as a solution algorithm, which is shown to be highly efficient for this class of programs.

Characterizations and Quantitative Estimation of Alkali-Activated Binder Paste from Microstructures

Alkali-activated binder (AAB) is recently being considered as a sustainable alternative to portland cement (PC) due to its low carbon dioxide emission and diversion of industrial wastes and by-products such as fly ash and slag from landfills. In order to comprehend the behavior of AAB, detailed knowledge on relations between microstructure and mechanical properties are important. To address the issue, a new approach to characterize hardened pastes of AAB containing fly ash as well as those containing fly ash and slag was adopted using scanning electron microscopy (SEM) and energy dispersive X-ray spectra microanalyses. The volume stoichiometries of the alkali activation reactions were used to estimate the quantities of the sodium aluminosilicate (N–A–S–H) and calcium silicate hydrate (CSH) produced by these reactions. The 3D plots of Si/Al, Na/Al and Ca/Si atom ratios given by the microanalyses were compared with the estimated quantities of CSH(S) to successfully determine the unique chemical compositions of the N–A–S–H and CSH(S) for ten different AAB at three different curing temperatures using a constrained nonlinear least squares optimization formulation by general algebraic modeling system. The results show that the theoretical and experimental quantities of N–A–S–H and CSH(S) were in close agreement with each other. The R2 values were 0.99 for both alkali-activated fly ash and alkali-activated slag binders.

Optimization Framework for Bicycle Network Design

This paper presents a new formulation for the network design problem as it relates to retrofitting existing roadway infrastructure for bicycles. The goal of the problem is, for a minimum cost, to connect all origin-destination pairs with paths where each roadway segment and intersection meets or exceeds a lower bound on its bicycling level of service. The length of each optimal path is constrained to be no greater than a given upper bound, which is expressed as a function of shortest path length. Experimental analysis on the Austin, Texas downtown region shows that a systems approach will yield different results than an approach that separately considers connecting each pair of origins and destinations, and that placing an upper bound on the amount of deviation from the shortest path will impact the design decisions. Model parameters, although the defaults are based on existing research, should be calibrated based on local data. Variants on the formulation are provided that allow for a trade-off between optimality and computational efficiency.

Centralized Carrier Collaboration Multihub Location Problem for Less-Than-Truckload Industry: Hybrid Hub-and-Spoke Network

A centralized carrier collaboration multihub location problem (CCCMLP) for the small to medium-sized less-than-truckload industry is addressed. In the CCCMLP, a central entity (e.g., a third-party logistics firm) seeks a set of collaborative consolidation transshipment hubs to establish a hybrid collaborative hub-and-spoke system that minimizes the total collaborative costs for the set of collaborating carriers. Previous studies focused on addressing the exchange of capacity without considering the location of transfer hubs and the routes that connect them. A carrier has the option either to collaborate or to ship its demand directly without collaborating. The decision depends on the expected profit margin over shipping directly while following a revenue-generating, rate-setting behavior. The CCCMLP was formulated as a variant of the P-hub location problem, which is NP-hard and solved with Lagrangian relaxation. Numerical experiments were conducted to gain insight into the performance of the CCCMLP formulation under various network sizes and numbers of hubs. The results indicate that larger expected profit margins from collaborative carriers applying revenue-generating behavior would increase the likelihood of collaboration by carriers. As the network size increases, the effect of hybrid hub location costs drops.