پژوهشنامه حمل و نقل
سال سیزدهم شماره 1 (پیاپی 46، بهار 1395)

Safety and ability of roads to pass vehicles with the minimum delays for drivers depend on good traffic order¡ and it can only be achieved if the signs as one of the most important means of traffic control¡ guide the drivers properly. Therefore¡ the signs must be so designed that legibility and visibility are supplied all day long. Traffic signs for example guide signs have special important for traffic control. Unfortunately¡ in many areas signs have been so designed and installed that driver expectations haven’t been considered. So in this project¡ by recognizing effective factors of legibility¡ variables were selected and they were used to build the regression model. In this model¡ legibility factors were considered as independent variables and driver expectations¡ based on the score as dependent variable were introduced. And finally¡ it was found¡ signs that were designed as the standard norms¡ more drivers pay attention to them.

One of the most important issues which transportation authorities are always confronted with is investment projects selection. After projects selection¡ due to resource constraints¡ selected projects require to be scheduled during the planning horizon. Integration of selecting and scheduling projects in a single model increases the accuracy of the results¡ but on the other hand increases complexity. In this paper¡ integrated model of selecting and scheduling transportation investment projects is introduced as a bi-level discrete network design problem. In this model¡ duration time and annual costs of candidate projects are unknown at first¡ but they will be determined by solving the model. A new measure to evaluate projects according to the threshold of user satisfaction is developed that in combination with the total travel time form the upper level objective function. Given the great complexity of the exact solution¡ a genetic algorithm (GA) is proposed to solve the model in a reasonable time. Also to optimize the GA parameters¡ the Taguchi method has been used. In order to evaluate the performance of the proposed algorithm¡ Sioux Falls urban transportation network is considered and three sample problems in different scales have been produced. The proposed GA and complete count method used to solve the problems. Numerical results show that proposed GA has an acceptable performance in both solution quality and solution time¡ especially in the large-scale problem.

The force-based formulation is attractive because it can model the spread of plasticity along the length of column using only one element and a number of integration points; however¡ their sensitivity to modeling parameters and models of material behavior has resulted in the estimation of residual displacement seismic demand to be important in this modeling method. In this research¡ the effects of two theories of lumped plasticity and distributed plasticity on modeling the concrete bridge column for estimating the residual displacement in near field zones were investigated. For estimation of residual displacement¡ the effects of reinforcing slide¡ geometrical parameters¡ and material behavior parameters were studied. Due to the importance of residual displacement in the performance evaluation of damaged structures¡ estimation of residual displacement is very crucial. In this study¡ the effects of different modeling methods as well as material behavior parameters on estimation of residual displacement were investigated. For this purpose¡ five different models were studied and the sensitivity of optimum model to various parameters was evaluated. The results showed that the theory of distributed plasticity for bridge concrete column has a good accuracy for estimating the residual displacement. Reloading strain parameters of concrete behavior and forced-based beam-column fiber element modeling parameters have an important role in estimation of residual displacement.

Diesel-electric locomotives are one of the foremost elements of transportation systems and they consume significant portion of transportation fuel usage. One of the effective parameters in fuel consumption of these vehicles is acceleration and braking strategy according to train speed profile. Although the available numerical methods for calculating the optimal speed profiles present acceptable results¡ the use of these methods in the calculation is too time-consuming. To reduce the computation time¡ using fuzzy logic as an efficient method has been suggested. In this paper¡ a fuzzy look-ahead control approach is implemented to optimize fuel consumption of diesel-electric locomotives. A fuzzy controller will decide the speed changing by considering the gradient of the path ahead. Hence¡ a speed profile close to optimal speed profile will be achieved. Furthermore¡ the performance of control algorithm has been improved by a sensitivity analysis of look-ahead controller parameters. The simulation results of developed algorithm show that at least 2.7 percents reduction in fuel consumption of diesel-electric locomotives can be achieved.

Evaluation of existing flexible pavement condition at network level requires structural assessment which is performed by FWD device using effective structural number (SNeff) parameter. Deflection data analysis requires layer thicknesses known which must be measured by GPR device or core drilling and pit boring. The AASHTO 93 selected as a base method¡ so FWD deflection data and layer thicknesses measured by GPR were converted to effective structural number. To reduce the cost and time consumption at network level assessment¡ layers thickness have been removed and a non-linear regression model was developed based on 3000 deflection basin parameters to determine effective structural number. Then the deflection basin parameters were studied and their correlation with SNeff determined using SPSS software. The analysis showed stronger relationship between AUPP¡ SCI and dimensionless parameters normalized by central deflection with SNeff. Thus a non-linear regression model obtained involving SCI¡ AUPP and D90/D0 with 0.9 determination coefficients. The developed model with 0.9 determination coefficients has considerable superiority in comparison with Jameson model with 0.84 determination coefficient which is previously presented.

Studying the marine accidents of Arvand channel and estuary during last five years¡ this research aims to locate the geographic place of risky points. Then through an annual and monthly chronological analysis and by application of time and place management¡ it tries to identify the eventful months in order to maximize the success of marine search and rescue measures. On the other hand¡ the chronological study of the mentioned accidents indicates that to what extent the accomplished attempts could have reduced or limited marine losses. As well¡ this study indicates that how marine rescue stations and the place of strategic search and rescue equipment suit the geographic place of eventful places. All accidents have been summarized into vessels'' marine accidents¡ borderline aggressions and crews'' illness and injuries. Then the factors of fatal and financial losses and damages have been identified based on which weighting is done. The weights of all accidents have been determined prior to chronological and spatial analysis. ARC VIEW GIS program has been applied for spatial analysis. The accidents position and their relevant intensity as the relevant specification have been entered into the program. The spatial analysis in this program has been done via Inverse Distance Weighting (IDW) method. The focus map of various accidents has been extracted. Place analyses in this research led to identification of 6 dangerous areas in view point of vessels accidents¡ 4 hazardous areas for human accidents and one highly risky area in view point of aggression to vessels. The annual time analysis show that during last 5 years¡ only the number of border aggressions has decreased while there is no obvious decreasing trend in vessels and human accidents. As well¡ the monthly time analysis shows 3 hazardous time sections in vessels accidents. The occurrence of accidents relevant to border aggressions is focused in the first quarter of each year. Management of rescue equipments - in two aspects of place and time - in hazardous areas as well as suggestion of having marine equipment in hazardous places is of the obtained results of place and time analyses in this research.

One of the most important factors in the failure of projects is the selection process of the delivery method. The wrong selection of the delivery method and consequently inappropriate selection of parties to the contract is a strategic mistake that causes bad consequences in projects. Identifying challenges in the delivery systems is very important in our country that causes significant economical revenues by attention and utilization in infrastructure projects which are factors of progress in the country. In this article¡ the most important factors of the three main delivery systems; including Design and Construction¡ Design-Bid-Build and Construction Management; in freeway projects are extracted and analyzed using questionnaire and SPSS software. Then¡ the best project delivery methods are identified and ranked by applying the analytical hierarchy process method.

The vehicle routing problem with time windows (VRPTW) is one of the most well-known combinatorial problems in the transportation domain. Because this problem belongs to NP-Hard problems¡ many researchers have presented meta-heuristics for it. In this paper¡ aimed at the disadvantages existed in the current Ant Colony System (ACS)¡ a modification of this algorithm called HACS is proposed for solving VRPTW. To improve the performance of ACS¡ two neighborhood search in the name of insert and swap algorithms are used. This new algorithm can avoid premature convergence and exploit more strong solutions. Finally¡ the effectiveness of the HACS on solving a number of the Solomon’s 56 VRPTW is validated by comparing the computational results with those previously presented in the literature. The results show that not only the proposed HACS algorithm can find good solutions but also seven best known solutions of the benchmark problem are also found by the proposed method.