This paper focuses on predicting the severity of freeway traffic accidents by employing twelve accident-related parameters in a genetic algorithm (GA), pattern search and artificial neural network (ANN) modelling methods. The models were developed using the input parameters of driver's age and gender, the use of a seat belt, the type and safety of a vehicle, weather conditions, road surface, speed ratio, crash time, crash type, collision type and traffic flow. The models were constructed based on 1000 of crashes in total that occurred during 2007 on the Tehran–Ghom Freeway due to the fact that the remaining records were not suitable for this study. The GA evaluated eleven equations to obtain the best one. Then, GA and PS methods were combined using the best GA equation. The neural network used multi-layer perceptron (MLP) architecture that consisted of a multi-layer feed-forward network with hidden sigmoid and linear output neurons that could also fit multi-dimensional mapping problems arbitrarily well. The ANN was applied during training, testing and validation and had 12 inputs, 25 neurons in the hidden layers and 3 neurons in the output layer. The best-fit model was selected according to the R-value, root mean square errors (RMSE), mean absolute errors (MAE) and the sum of square error (SSE). The highest R-value was obtained for the ANN around 0.87, demonstrating that the ANN provided the best prediction. The combination of GA and PS methods allowed for various prediction rankings ranging from linear relationships to complex equations. The advantage of these models is improving themselves adding new data.

There is a constant effort to improve the performance of asphalt-concrete (AC) mixtures. Among various modifiers for asphalt, fibers have received much attention for their improving effects. This paper introduces the novel concept of hybrid reinforcement of AC mixtures using polypropylene (PP) and glass fibers. Individually, glass fiber reinforced AC and PP fiber modified AC mixtures have exhibited superior performance compared to other fiber reinforced samples. Therefore, in this work, these two types of fibers were used simultaneously to improve the performance of the AC mixtures. This type of hybrid AC composite can be engineered by taking advantage of the tacky property of PP fiber around its melting point and the high modulus of glass fiber. In this way, PP fibers with the length of 12 mm were blended with bitumen at different percentages. Glass fibers with the length of 12 mm were also added to aggregates. Marshall and Specific Gravity tests were performed on hybrid reinforced asphalt-concrete (HRAC) samples by taking advantage of a Superpave Gyratory Compactor. In the case of the bituminous specimens, penetration, softening point and ductility tests were carried out. The results revealed that PP fibers decrease penetration and ductility of modified bitumen, while the softening point value is increased compared to unmodified bitumen specimen. Marshall Test results illustrate that PP can statistically affect the properties and improve the consistency of the mixture. Using a combination of 0.1% of glass fiber plus 6% of PP presented the best hybrid reinforcement through increasing stability and decreasing flow. Therefore, it is concluded that this novel HRAC is suitable for use in hot regions due to growth in the void total mix (VTM) and stability.

This paper compares two fuzzy clustering algorithms – fuzzy subtractive clustering and fuzzy C-means clustering – to a multi-layer perceptron neural network for their ability to predict the severity of crash injuries and to estimate the response time on the traffic crash data. Four clustering algorithms – hierarchical, K-means, subtractive clustering, and fuzzy C-means clustering – were used to obtain the optimum number of clusters based on the mean silhouette coefficient and R-value before applying the fuzzy clustering algorithms. The best-fit algorithms were selected according to two criteria: precision (root mean square, R-value, mean absolute errors, and sum of square error) and response time (t). The highest R-value was obtained for the multi-layer perceptron (0.89), demonstrating that the multi-layer perceptron had a high precision in traffic crash prediction among the prediction models, and that it was stable even in the presence of outliers and overlapping data. Meanwhile, in comparison with other prediction models, fuzzy subtractive clustering provided the lowest value for response time (0.284 second), 9.28 times faster than the time of multi-layer perceptron, meaning that it could lead to developing an on-line system for processing data from detectors and/or a real-time traffic database. The model can be extended through improvements based on additional data through induction procedure.

Modified Augmented Lagrangian Genetic Algorithm (ALGA) and Quadratic Penalty Function Genetic Algo­rithm (QPGA) optimization methods are proposed to obtain truss structures with minimum structural weight using both continuous and discrete design variables. To achieve robust solutions, Compressed Sparse Row (CSR) with reordering of Cholesky factorization and Moore Penrose Pseudoinverse are used in case of non-singular and singular stiffness matrix, respectively. The efficiency of the proposed nonlinear optimization methods is demonstrated on several practical exam­ples. The results obtained from the Pratt truss bridge show that the optimum design solution using discrete parameters is 21% lighter than the traditional design with uniform cross sections. Similarly, the results obtained from the 57-bar planar tower truss indicate that the proposed design method using continuous and discrete design parameters can be up to 29% and 9% lighter than traditional design solutions, respectively. Through sensitivity analysis, it is shown that the proposed methodology is robust and leads to significant improvements in convergence rates, which should prove useful in large-scale applications.

No abstract