Abstract:Based on the Logit choice model and elastic demand function， a Logit-based stochastic ridesharing user equilibrium model with elastic demand was built， and a self-adaptive parallel projection algorithm was proposed to solve the ridesharing traffic assignment problem. Finally， the effectiveness of the proposed model and algorithm was verified by two benchmark urban transportation networks. It is shown that the elastic demand parameter and user perception dispersion parameter are critical factors affecting the ridesharing， and the travel demand and average travel time decrease with the increase of the elastic demand parameter and user perception dispersion parameter. The proposed algorithm can solve the large-scale ridesharing traffic assignment problem efficiently and obtain the globally optional solution.

Abstract:Based on a crossed structure， a service strategy for passengers to visit alternative hubs in turn under the failure of the hub and imperfect information was proposed， and a reliable hierarchical location model with the capacity limitation was established. The example analysis shows that the model can enhance the reliability of the location planning and reduce the penalty cost due to the failure of the hub. Furthermore， through the sensitivity analysis of model parameters， it is found that for the transport hub network with high failure probability， increasing the number of hubs alone can not improve the reliability of the location planning， while the transport hub network with high reliability has greater flexibility to deal with events that may lead to high losses.

Abstract:From the perspective of methodology， a traffic emergency management method under public health emergencies was proposed. The first stage was to build a traffic flow distribution model and the second stage was to build a bi-level traffic management model. Numerical simulation results show that by adjusting the ticket price of high-speed railway， optimizing the average number of passengers carried per private vehicle and the ridesharing subsidy scheme， the total number of people from high social contact frequency areas or in high risk travel modes can be decreased effectively.

Abstract:A traffic assignment model for the priority of emergency logistics was proposed， and a corresponding numerical method was designed. In a general transportation network， the route choice of emergency logistics follows the Cournot-Nash （CN） principle in which emergency logistics minimizes the overall cost/time of emergency logistics. Other social vehicles respect the priority of emergency vehicles， and then follow the Wardrop user equilibrium （UE） principle aiming at minimizing their own travel cost， by giving prior right of traveling to emergency vehicles. The proposed UE-CN mixed traffic assignment model may not have unique solution， and an algorithm was designed for solving the model. Finally， the effectiveness of the model and algorithm was verified through numerical examples.

Abstract:By constructing an evolutionary game model under public participation， how to avoid the opportunistic behavior in risk-sharing renegotiation for transportation infrastructure public-private partnership （PPP） projects was analyzed. It is shown that public participation can force enterprises to adopt non-opportunistic behavior strategy when the benefit of non-opportunistic behavior is less than that of opportunistic behavior； the increase of punishment level for the opportunistic behavior of enterprises is an important way for the avoidance of opportunistic behavior under the low probability of public participation；a more severe goverment punishment should be taken to chang enterprises’ opportunistic behavior under the low probability of public participation because the different probabilities of public participation have different requirements for government punishment.

Abstract:Six specimens of high-performance steel beams made of Q460 and Q500 steel were designed to study the overall stability in bending. Considering the initial defects of the specimens， a finite element model was established via Abaqus to simulate the test process. Finally， the test and finite element simulation results were compared with the relevant provisions of the Standard for Design of Steel Structures （GB 50017―2017） and the Specifications for Design of Highway Steel Bridge （JTG D64―2015）. It is shown that the simulated load-displacement curve is in good agreement with the test results； the applicability of the two codes on the overall stability of Q460 and Q500 high-performance steel beams is verified.

Abstract:Four-point flexural tests were carried out on 14 reinforced high-strength engineered cementitious composite （HS-ECC） beams. The effects of concrete types， reinforcement ratios of longitudinal bars and stirrups on the flexural and shear behaviors of reinforced beams were explored. Then， the bending capacitity of the HS-ECC beams with stirrups was calculated based on the plane-section assumption and material constitutive models， and the shear capacitiy of HS-ECC beams without stirrups was calculated according to the specifications at home and abroad. Finally， the finite element model of HS-ECC beams was established via Abaqus. It is shown that： all the HS-ECC specimens with stirrups are flexural failure； the ultimate flexural load and the stiffness of HS-ECC beams increase with the reinforcement ratio， but the ductility remains stable； HS-ECC beams with stirrups demonstrate superior flexural performance and crack distribution capacity compared with conventional concrete beams； the failure modes of HS-ECC beams without stirrups vary from flexural failure to shear failure with the increasing reinforcement ratio； the higher the reinforcement ratio of HS-ECC beams， the higher the shear capacity and stiffness， while the lower the ductility； calculation results of the bending capacity of the HS-ECC beams are in good agreement with experimental values； the load-displacement curves of the HS-ECC beams are simulated with the finite element model effectively.

Abstract:Four single-stud and two grouped-stud fatigue push-out tests were conducted. The fatigue test results of studs were compared with the stud fatigue S-N curve in domestic and international codes， and the finite element analysis based on material plastic damage model was done. It is shown that： the stud root fracture and local UHPC crush around the stud root are the main failure mode； compared with single-stud specimen， the fatigue life of grouped-stud specimen is reduced by 26.9%， and the shear stiffness of grouped-stud specimen degrades more rapidly during cyclic loading； particularly， the stud shear stiffness ratio decreases to 16% when loading cycles reach 12.0% of fatigue life； the AASHTO based evaluation has the largest safety redundancy among the codes while the JSCE based evaluation is closest to the test results but with a low level of safety redundancy. In addition， the finite element analysis results show that the maximum cyclic strain amplitude at stud root obtained from grouped-stud model is larger than that obtained from single-stud model， and the plastic accumulation damage of UHPC around stud root in grouped-stud model is severer than that in single-stud model.

Abstract:By substituting the orthotropic steel desk（OSD） of nearside lane and emergency lane with a ultra-high performance concrete（UHPC） waffle deck for a large-span bridge， a hybrid deck system was formed. Thus，a transverse connection detail to link the two types of decks was proposed. Specifically， positive and negative bending specimens of the transverse connection detail were designed and tested to determine the failure mode， ultimate bearing capacity， anti-cracking performance， and cooperative moment bearing performance. Lastly， a method of calculating the bearing capacity after cracking was presented by taking into account the residual stress of UHPC. The bridge decks on both sides are able to cooperate with each other in bearing the external load through the transverse connection detail， indicating better plastic deformation capability. The OSD and UHPC waffle decks conform to the plane section assumption in elastic stage. With the proposed bearing capacity calculation method， the bearing capacity of the positive bending specimen can be predicted with satisfactory accuracy and the error is within 8%. Local buckling is an important factor in determining the ultimate bearing capacity of the negative bending specimen， which should be considered when calculating the bearing capacity.

Abstract:Uniaxial tensile tests and high stress repeated tension-compression tests for 32 specimens of sleeve grouting lap connectors named APC connector were conducted to study the effect of lap length on the mechanical properties of APC connector. It is shown that： after the high stress repeated tension and compression， the bearing capacity of the specimens is improved， while the ductility of the specimens is reduced due to the further development of the cracks between the steel bar and the grout； the residual deformation u₀ （single tension） and u₂₀ （high stress repeated tension and compression） of the specimens are reduced as the lap length of the specimens is increased； with the increase of lap length， the longitudinal compressive strain and hoop tensile strain of the middle section of the sleeve near the steel bar are reduced under the ultimate load state when the specimens are stretched under uniaxial tension and the last tensile part after repeated tension-compression loading； based on the gray relation analysis， the steel content of the connector has the greatest correlation with the ultimate bond strength. Finally， the ultimate bond strength and critical lap length formulas were put forward with the introduction of the grouting defect coefficient ω.

Abstract:A method of cone penetration test （CPT） combined shear wave velocity （V_s） test （CPT-V_s combination method） was proposed based on logistic regression algorithm. In connection with the liquefaction cases of Tangshan earthquake， the comparison was done among CPT-V_s correlation method， q_c/G₀ method and CPT-V_s combination method. It is shown that the site liquefaction discrimination rates of CPT-V_s correlation method， q_c/G₀ method and CPT-V_s combination method are 89%， 78% and 100%， respectively， and the overall accuracies of liquefaction discrimination of the three methods are 94%， 50%， and 94%， respectively； the application scope of CPT-V_s correlation method is limited by parameter values， and the q_c/G₀ method is too conservative， resulting in a large discrimination error in the non-liquefied area； the method proposed is not affected by the limitation of parameter values， and can give satisfactory results.

Abstract:An evaluation framework based on five evaluation indicators was proposed. The Pearl growth curve function and negative exponential function were selected to standardize the positive and negative indicators， respectively. Then， the expert scoring method and analytic hierarchy process were used to determine the weight of each indicator. Finally， the integration degree model of autonomous driving road testing was built. The case study， based on the measured data and simulation results of automatic driving road testing under two types of urban road scenarios and one type of highway scenario in Shanghai， is conducted to verify the validity and effectiveness of the integration degree model. The results show that the risk-avoiding disengagement frequency is the most critical evaluation indicator which characterizes the integration degree of autonomous vehicles into existing road traffic system under both urban road scenarios and highway scenario； the integration degree under the highway scenario is significantly higher than that under the urban road scenario； increasing the testing mileage， the testing duration and the complexity of testing scenarios can promote the maturity of autonomous vehicle technology.

Abstract:Multi-dimensional driving behavior data， including driver characteristics， road environment， and vehicle operation variables， were collected based on natural driving experiments. Then， the standard database of aggressive driving behavior was created after data cleansing and filtering. The significance analysis was used to obtain useful indexes， with which an eight-dimensional index set for aggressive driving behavior prediction was built. Finally， a two-layer time series model for aggressive driving behavior prediction was constructed. The first layer of the model is an artificial neural network. The second layer of the model is a long short-term memory （LSTM） network with an attention mechanism module. It is shown that the proposed model can increase the prediction accuracy by 10%；the two-layer structure and attention mechanism have a good improvement for the prediction accuracy （5% and 3%， respectively）.

Abstract:Based on the driving simulation experiment， four variable message sign （VMS） frontal distances of 3.0 km， 2.5 km， 1.5 km， and 0.7 km were set， and the steering wheel， speed and position data of 32 drivers were collected during the process of changing lane and leaving freeway after the recognition of VMS with four frontal distances. Then， the steering behavior， lane changing behavior， and deceleration behavior during the lane changing process were analyzed. It is shown that the VMS recognition procedure can be classified as the multi-tasking driving behavior， because drivers have to complete VMS recognition， route diversion decision， deceleration and lane changing in a short time； if the VMS frontal distance is not long enough， drivers should turn the steering wheel quickly and sharply so as to change lanes continuously and urgently； to successfully enter the deceleration lane， some drivers reduce speed during the lane changing process， which increases the traffic risk； when the VMS frontal distance is too long， the short-term memory effect on VMS increases the driving load of drivers.

Abstract:In order to improve the image segmentation accuracy for cohesive aggregates， a image segmentation model （Multi-ResUnet model）based on Inception network and residual connection optimization was proposed. The three-dimensional aggregate characterization system V3.0 developed by the laboratory itself was used to collect the images of cohesive aggregates and the sample set of the image segmentation model was created. Then， the image segmentation model was used to train the sample set. It is shown that compared with the Watershed algorithm and Unet model，the image segmentation model improves 30.46% and 2.11% in precision， 4.68% and 1.85% in recall， and 25.95% and 2.47% in accuracy， respectively.

Abstract:In view of the characteristics of multi-variety， small batches and low automation degree， the sewing production line balancing problem was addressed considering worker moving cost of multi-process operation and adding an appropriate amount of equipment to bottleneck processes. A dual-objective optimization model was constructed to maximize balance rate and minimize takt time by introducing the concept of virtual workstation and standard work-in-process （SWIP） inventory. Finally， the enumeration algorithm and multi-objective genetic algorithm based on greedy search strategy were designed to obtain newly added equipment scheme， SWIP quantity and line balancing optimization scheme. A case study of sewing production line was carried out. It is shown that the line balancing rate and production efficiency are greatly improved. The computational results demonstrate the rationality and effectiveness of the proposed model and algorithm.

Abstract:Taking the extreme driving conditions as negative stimulus， the subjective feelings of 12 subjects as front-seat passenger were obtained in the process of real-vehicle testing by means of evaluation scale and interview， and the skin electrical activity data were recorded by the physical polygraph. On this basis， the change of the passengers’ subjective evaluation was studied by comparing the data before and after experiencing the unintended vehicle maneuver and using the hypothesis testing method. The results show that the negative stimulus improves the passengers’ expectation of vehicle control ability， but increases their anxiety about the uncertainty of vehicle decision-making intention. In addition， the combination of the data of skin conductance level and the data of the event during the test can reflect the change of passenger’s expectation.