For the following reasons, most LIS platforms only offer the information on trucks and cargoes, failing to support rapid T-C matching: Despite the mammoth demand, the highway freight market is highly fragmented, with information asymmetry between numerous small carriers; it generally takes a long time to complete T-C matching, and the highway service has a poor reputation; truck suppliers rarely cooperate with each other, making it difficult to optimize the allocation of social resources.
To solve these problems in highway freight industry, this paper attempts to explore the relevant issues of highway T-C matching.
Literature review
.1. Previous research on D-P transport
Drop-and-pull transportation refers to a transportation mode wherein a truck pulls the trailer to its destination; drops the trailer, and hooks up a new trailer to be transported to its destination.
Existing studies on drop-and-pull transportation are mainly focused on drop-and-pull vehicle safety, route optimization and vehicle performance. In terms of drop-and-pull vehicle safety, Godbole et al. [1] studied how parameters affected the amount of the dynamic loading of the drop-and-pull vehicles, including vehicle loading level, center of mass position, suspension setting of the truck-trailer, and amount of damping of vibration components. The influence of these parameters major on vertical motion of trailer chassis. Salati et al. [2] experimented wind tunnel on heavy duty trucks equipped with front-rear axle trailer allocation equipment and studied the aerodynamic drag of European heavy duty trucks. In the improvement of drop-and-pull route optimization, Wang and Chan [3] developed a mathematical model in the form of multi-objective integer programming model to obtain the optimal number of vehicles that will follow the most effective routes for the pick-up and delivery of multiple products or components along the production flow, which will minimize energy consumption and operational cost. Mirmohammadsadeghi and Ahmed [4] did a research focuses on random demand for truck and trailer routing problem (TTRP) with time Windows. In terms of vehicle performance, Pflug [5] used different combinations of trucks and trailers, and utilized 3D simulations to analyze changes in vehicle lateral stability, damping performance and vibration frequency, under extreme driving conditions. Kim [6] learned the mechanism of drag reduction by studying the vehicle model precursor. The research results are expected to provide useful information for the design of new CRF models and the improvement of aerodynamic performance of heavy vehicles.
Previous research on matching
Graph matching is a hot topic among researchers engaging in matching. Chen et al. [7] designed an effective singularity method for graphic matching, which searches for multiple smooth curves at each detection point by checking path smoothness. Based on geometric graphs, Dwivedi and Singh [8] proposed an error-tolerant graph matching method: the vertex distance and edge distance between two geometric graphs were introduced to compute the distance between the graphs, and relied on the computed distance to match the error-tolerant graphs. Arhid et al. [9] solved the problems of shape matching and retrieval, and developed a novel method based on local matching, in which each 3D object is broken down to its basic elements, and shape descriptors are derived from the elements to compare the similar points.
Matching techniques have been applied to marriage matching and person-post matching. On marriage matching, Boudreau and Knoblauch [10] presented a marriage matching mechanism that maximizes the welfare, and held that the mechanism could be promoted to various welfare measures. Hoefer and Wagner [11] investigated two-sided matching markets with locality of information and control, focusing on locally stable marriage with strict preferences. On person-post matching, Miranda et al. [12] created a competence management model based on ontology, which represents competences that support a wide range of scenarios, enables interoperability and cooperation among different and heterogeneous tools, and execute queries and inference operations over these competences, thereby realizing long-term accurate matching between persons and posts. Brandmeier et al. [13] designed a competence management system that can identify various competences and enhance the ability of competence retrieval. To realize accurate person-post matching, Enăchescu [14] established a rather complete knowledge representation diagram for e-recruitment ontologies, and set up a job recommendation system based on knowledge ontology, which can update personnel and job information easily and rapidly.
Many scholars have summarized the matching methods and utilized new methods for empirical analysis. Kim et al. [15] simplified the linear feature matching method using decision tree analysis, weighted linear directional mean, and topological relationships, and proved that the simplified method is highly robust. Many other matching methods have emerged, including tabu algorithm [16], semantic matching [17], neural network (NN) [18], and particle swarm optimization (PSO) [19]. These methods promote the intelligent T-C matching and multi-index evaluation of truck provider and cargo supplier.
