Research of Intersection Groups Optimization Control Method Based on Critical Path Identification

. An optimal intersection groups control method based on critical path identification is proposed. The control algorithm uses a binary search to gradually determine the scope of the intersection groups, uses the duality method to express the intersection groups, uses breadth-first search algorithm to solve the critical path, and finally uses the branch-and-bound method to solve the lane canalization, realizes the optimized output of phase-sequence and timing of intersection control signals. The control algorithm is realized using C++ language, including the following functional modules: intersection range dynamic definition, critical path identification, space-time resource optimization, and online timing adjustment of signal timing parameters and so on. Finally, the control algorithm is verified by the actual road network of Changzhou City of China, the result shows that, the efficiency of traffic operation is significantly improved.


Introduction
The intersection groups of urban is a set of intersections of the city road networks that are geographically adjacent to each other, and there is a strong correlation [1]. Solving urban traffic congestion from the perspective of the intersection groups has been recognized by most scholars. In 1978, Turner I.F. and Shannon G.F. first mentioned the concept of intersection groups under few road interactions and applied it to cooperative control among three intersections [2]. Liu [3] (2011) studied and analyzed the process of interaction between traffic flow and intersections under the vehicle queuing variation of trunk road and control signals at intersections, based on this, a multi-lane oversaturated signal cooperative control system was proposed. Yang Xiaoguang first proposed the concept of intersection groups signal control in 2001 [4] and conducted a systematic study of urban traffic congestion problems. Most scholars have carried out research on traffic conditions, optimization objectives, control strategies, and control methods, however, they did not consider the issue from the perspective of system design and software implementation [5][6][7][8][9][10]. This article constructed the intersection groups control system, optimized the algorithm of each subsystem, and finally realized the whole system using C++.

The framework of intersection groups priority control system
The intersection groups optimization control system is divided into several modules: Intersection Groups scope division, critical path identification, space-time resource optimization, timing parameters online adjustment, data transmission. The specific control framework and data flow are shown in Figure 1.

Intersection groups scope division
The intersection groups scope division module, firstly calculates the coordination control coefficient CF and the flow imbalance coefficient IF between the intersections based on traffic, period, average vehicle speed, and section length of intersections, and use the sum of CF and IF as the road segment correlation value. The self-organizing neural network adopts fuzzy means clustering algorithm to classify the strength of association. Using the branch-andbound method to gradually reduce the scale of the road networks to determine the signal control range of the intersection groups.

Critical route identification
First, the junction map of the intersection groups is converted into a dual graph. The breadth-first search algorithm is used to search all the nodes in the dual graph, that is, all the road sections in the node graph. By comparing the ratio of the number of vehicles from the starting point to the end point, the degree of linkage discreteness index I 1 is calculated, and the degree of correlation between the length of the functional area and the distance of the road sections are used to calculate the degree of blocking-relevance index I 2 , and obtains the degree of path correlation I after dimensionless processing. By comparing the path association degree values of each path, the effective path with the largest path association degree value is determined as the key path of the current intersection groups. The value of the path association degree I is:

Space-Time resource optimization
The mathematical programming method is used to realize the comprehensive optimization of time-space resources in intersection groups: the decision variables are represented by algebraic symbols, a binary mixed integer linear programming model is established according to the constraints and optimization objectives. The branch-andbound method is used to solve the canalization design, phase design and timing control parameter design scheme for each intersection of the intersection groups. There are nine phase-sequence options available after the design is implemented, one of which is shown in the figure.

Signal timing parameters online adjustment
Timing parameters online adjustment module consists of two sub-modules: green signal ratio adjustment module and phase difference optimization module. The green signal ratio optimization module receives the real-time detection data, and the module continuously reads 10 pieces of data before the predetermined phase ends, each piece of data represents the detection data of each time unit in the detection period of 10 seconds. If extended green light time plus the minimum phase time is greater than the maximum green light duration, then the algorithm is extended to the maximum green light duration, and it switches to the next phase.

Development platform
The software development platform is selected as follows: • Development language: C++; • Database management system: Oracle 9i or later.

Data description
The interface between independent modules uses data interface instead of Socket or other interface modes. The client program is designed based on the QT graphical interface library and provides a friendly man-machine interface. The database structure is shown in Table 1.

Graphical interface design
The graphical interface module needs to display information: intersections input information, road section information, road network topology, and other information calculated by the software, such as target intersections, intersection groups, and critical paths. Figure 3 shows the software's main graphical interface. The left side shows the road network structure and the bottom map. In the figure, the solid dots in the figure represent the intersections, the red dots are the target intersections, the blue dots are the intersections within the intersection groups, and the red lines are the critical path. The right side of the figure indicates the intersection number, traffic capacity, phase sequence information and so on. The information of the corresponding intersection and road section can be displayed according to the selection taken by the mouse.

Conclusion
In this paper, an optimal control method for intersection groups based on critical path identification is proposed.
The application results verify that, it can effectively improve the existing traffic operation status. The next step of the research is to add a bus priority control and special vehicle guidance module, to provide more reference for engineering practice.