CN106197455B

CN106197455B - A kind of real-time dynamic multipath mouth path navigation quantum searching method of urban road network

Info

Publication number: CN106197455B
Application number: CN201610605008.XA
Authority: CN
Inventors: 胡文斌; 聂聪; 邱振宇; 杜博; 王欢; 严丽平
Original assignee: Wuhan University WHU
Current assignee: Wuhan University WHU
Priority date: 2016-07-28
Filing date: 2016-07-28
Publication date: 2018-11-13
Anticipated expiration: 2036-07-28
Also published as: CN106197455A

Abstract

The invention discloses a kind of real-time dynamic multipath mouth path navigation quantum searching methods of urban road network, the value at cost for the generation that influences each other when the preference value of route influence generation and vehicle operation is combined using road itself and forms comprehensive assessment index value of utility, the quality of path navigation scheme is assessed using the size of value of utility, and using the value of utility of all path navigation schemes of quantum calculation parallel computation, go out satisfactory path navigation scheme using quantum searching effective search.The present invention has fully considered the various factors for influencing the coast is clear, and quantify finally integration to the influence degree of traffic by various factors and obtain value of utility, uses the quality of value of utility accurate judgement path navigation scheme.Introduce quantum calculation and quantum searching simultaneously so that the result of calculation of value of utility can be obtained in real time, and thus obtain suitable path navigation scheme, under the premise of meeting each driver's individual interest so that the traffic congestion of entire city road network is obviously improved.

Description

Urban traffic road network real-time dynamic multi-intersection path navigation quantum search method

Technical Field

The invention belongs to the technical field of computer science and intelligent traffic systems, and particularly relates to a real-time dynamic multi-intersection path navigation quantum search method for an urban traffic network.

Background

The urban traffic network in large and medium cities is increasingly congested, time cost, management cost and economic cost caused by congestion are increasingly high, the travel time of residents is increased due to traffic congestion, the working efficiency and the life quality of people are affected, urban development is restricted, energy consumption and tail gas emission are increased, environmental pollution is aggravated, and the problem of congestion of the urban traffic network is solved, so that the nation and the people are benefited. However, the urban traffic network structure is difficult to change, the road resources are limited, and efficient path navigation and reasonable road resource allocation become main ways for solving the urban network congestion.

The path navigation can be divided into static path navigation and dynamic path navigation, wherein the static path navigation refers to seeking the shortest path by taking conditions such as physical geographic information, traffic rules and the like as constraints, and the dynamic path navigation is to combine real-time traffic information to timely adjust a pre-planned optimal driving route on the basis of the static path navigation until the optimal path is finally obtained when the optimal driving route reaches a destination. At present, most of mature path navigation systems applied to the market are based on static path navigation, mainly including Dijkstra algorithm, Lee algorithm, Floyd algorithm, blind search, a-x heuristic algorithm and the like, but users are not satisfied with the existing system in the face of traffic reality with a lot of unstable factors. Although static path navigation can quickly find the optimal path of a single vehicle, local road congestion and relative idle of other local resources are difficult to avoid due to lack of coordination among vehicles, and when a traffic accident and a traffic jam occur, the static path navigation cannot timely change a route according to road condition information in real time. Providing real-time dynamic path navigation for vehicles is therefore critical to relieving road traffic congestion. The vehicle dynamic path navigation predicts the future traffic flow based on historical and current traffic information data and is used for timely adjusting and updating the optimal driving route, thereby effectively reducing road blockage and traffic accidents. The importance of traffic prediction in dynamic path navigation is gradually highlighted, and more researchers use a Kalman filtering method, a time sequence method, a neural network method, Markov prediction, grey prediction theory and the like to carry out deep research on traffic information prediction. Although it is not difficult to provide real-time path navigation information for vehicles with the rapid development of networks, the accuracy of a real-time prediction model is limited by a simple dynamic real-time prediction flow model, so that the processing capability of real-time traffic emergency is poor, the complex dynamic real-time prediction model has a plurality of consideration factors, and the calculation complexity is exponentially increased along with the increase of the road network scale, so that the current dynamic path navigation is not mature, and mostly stays in a theoretical stage. The contradiction between the accuracy and the complexity of the dynamic real-time prediction model limits the development of dynamic path navigation.

Disclosure of Invention

In order to solve the technical problems, the invention provides a real-time dynamic multi-intersection path navigation quantum search method for an urban traffic network, which is used for evaluating a path navigation scheme by comprehensively considering various factors influencing traffic and quantizing the factors to obtain the navigation scheme capable of effectively relieving traffic jam and maximizing the utilization rate of road resources of the urban traffic network.

The technical scheme adopted by the invention is as follows: a real road network real-time dynamic multi-intersection path navigation quantum search method maps a real road network into a model graph R (B, E), wherein B represents an intersection node set, and B represents an intersection node set_i(i 1, 2.., r) represents a single intersection node, r is the total intersection number, and E represents a set of road segments with directions; suppose there are n vehicles in the road network, and any vehicle w has a current starting point P_sAnd destination endpoint P_dThen a feasible path of the vehicle is represented by the nodes of the continuous adjacent intersections as { P_s,...,P_i,...,P_d}; each vehicle selects a feasible path, and the driving paths of all vehicles form a feasible path set FPS_nI.e. a path navigation scheme;

characterized in that the method comprises the following steps:

step 1: initializing a vehicle set { v) according to the number n of vehicles, start and stop point information and an optional path of each vehicle₁,v₂,...,v_nAnd optional Path setWherein v is_iIt means that the (i) th vehicle,an alternative path representing the ith vehicle;

step 2: for vehicles and their alternative paths 0,1_iCarry out quantum coding { |0>,|1>,...,|2^n×h-1>That the quantum states are determined to represent all completelyA path navigation scheme; wherein b is_iRepresenting the number of selectable paths of the ith vehicle, and h representing the minimum number of binary digits required for encoding the selectable paths;

step 3, determining independent multiplication factors α of each influence factor according to road condition information_i,β_jDetermining a utility value calculation function U (x); each path navigation scheme corresponds to an independent variable x value;

and 4, step 4: preparing an equal-weight superposition state | x > of the path navigation scheme, and calculating a utility value | U (x) | corresponding to each path navigation scheme x to obtain an equal-weight superposition state | U (x) >, of a utility value function;

and 5: determining an empirical value k of utility value, equal weight superposition state | U (x) of utility value function>Carrying out quantum search to search out utility value | U meeting the requirement_s>；

Step 6: outputting utility value U meeting the requirement_sAnd a corresponding path navigation scheme for performing path navigation on each vehicle.

Preferably, the utility value function U (x) in step 3 is:

U(x)＝Fr(x)×(α1×Rs(x)+α2×Sl(x)+α3×Ls(x)+α4×Os(x)+α5×Fd(x))-(β1×Ta(x)+β2×Tc(x)+β3×De(x)+β4×Oc(x)+β5×Tl(x))

wherein Fr (x) represents whether the road section can be reached, 1 represents reachable, and 0 represents unreachable; rs (x) represents the road section condition and takes the value of [0, 1%](ii) a Sl (x) represents a speed limit and takes the value [0, 1%](ii) a Ls (x) represents the illumination condition of the road section and takes the value of [0, 1%](ii) a Os (x) represents driver compliance with system recommendations, and takes the value [0, 1%](ii) a Fd (x) represents the familiarity of the driver with the road section and takes the value of [0, 1%](ii) a Ta (x) represents road influence caused by sudden traffic accidents or temporary control and the like, and takes a value of [0, 1%](ii) a Tc (x) represents the time cost spent by the selected path, taking the values [0, ∞](ii) a De (x) represents the distance cost consumed by the selected path and takes the values of [0, ∞](ii) a And Oc (x) represents the oil consumption cost of the selected path and takes the values of [0, ∞ ]](ii) a Tl (x) represents the influence of traffic lights and takes the value [0, 1%]；α_i(i＝1,2,...,5)、β_iEach of (i ═ 1, 2., 5) represents an independent multiplication factor corresponding to each influence factor.

Preferably, the specific implementation of step 4 comprises the following sub-steps:

step 4.1: quantum equal-weight superposition state for preparing initial independent variable path navigation scheme by utilizing Hadamard gateWherein N represents the total number of quantum states;

step 4.2: unitary conversion circuit U corresponding to design function_U(x)And an auxiliary qubit | z usable for implementing a functional computation>；

Step 4.3: inputting an equal-weight superposition state of a path navigation scheme, and calculating a function U (x) in parallel:

step 4.4: and obtaining the equal-weight superposition state | U (x) >, of the utility value function.

Preferably, the specific implementation of step 5 comprises the following sub-steps:

step 5.1: giving an oracle function f (y) for determining the target state and setting the corresponding quantum wire;

after the equal weight superposition state | U (x) > of the utility value function is judged by the oracle function, the state with the function value f (x) of 1 is the target state;

step 5.2: accumulating the target states to obtain a target state number m and calculating a comprehensive utility value target state | U_a>；

Wherein, a_iRepresents the target state, | a_i>A quantum form representing a target state;

step 5.3: according to | U_a>Determining an oracle query O, determining an O transform;

O＝I-2|U_a><U_a|；

wherein I represents AND | U_a>The equal weight superposition states with the same quantum bit number,<U_ai represents U_a>The conjugate vector of (a);

step 5.4: according to the equal weight superposition stateDetermining a D transform;

wherein,is an equally weighted stack of all the basic states,h represents Hadamard transform, used for preparing equal weight superposition state,representing the preparation of an n x h bit equal weight superposition state; n represents the total number of quantum states, | i>Represents the ith quantum state;

step 5.5: determining G-transform G-DO once through O-transform and D-transform;

step 5.6: equal weight superposition state | U (x) to utility value function>To carry outSecond G transform, round represents the nearest integer;

step 5.7: utility value state | U of observed output_out>And a path navigation scheme | x corresponding thereto_out>Searching out utility value | U meeting the requirement in time limit_s>；

Step 5.8: output utility value state | U_s>Corresponding path navigation scheme x_sThe navigation path selected for each vehicle.

Preferably, the specific implementation of step 5.7 comprises the following sub-steps:

step 5.7.1: observing the output after G transformation is completed to obtain utility value U_outAnd when the current search has been used t_s；

Step 5.7.2: if t is_s<t_maxThen the following step 5.7.3 is performed, where t_maxRepresenting a maximum navigation time interval that can guarantee the real-time performance of path navigation; otherwise, the following step 5.7.5 is performed;

step 5.7.3: if U is present_out<k, then t_s＝t_s+t_cAnd go back to perform said step 5.7.2, where t_cRepresents the time required to perform the RGQS method once; otherwise, the following step 5.7.4 is performed;

step 5.7.4: fruit of Luo Guo U_out<k_mThen k is equal to U_out，t_s＝t_s+t_cAnd go back to perform said step 5.7.2, where k_mIndicating an ideal utility value set empirically; otherwise, the following step 5.7.5 is performed;

step 5.7.5: u shape_s＝U_outOutput U_s。

The invention constructs a real-time dynamic multi-intersection traffic model of an urban traffic network, integrates various influence factors of urban traffic into utility values to evaluate the quality of a path navigation scheme; quantum computation and quantum search are introduced to solve the problem of real-time computation and search of utility values, after initial road conditions are determined, the algorithm provided by the invention can be used for real-time computation and search to obtain appropriate utility values and corresponding appropriate path navigation schemes, and path navigation is provided for all vehicles, so that traffic of the whole city is effectively relieved, and the utilization rate of urban road traffic resources is maximized.

Drawings

Fig. 1 is a real road network and model map according to an embodiment of the present invention.

Fig. 2 is a flow chart of the RGQS method in an embodiment of the present invention.

Fig. 3 is a flowchart of the UVCQC algorithm according to an embodiment of the present invention.

Fig. 4 is a schematic diagram of a UVCQC algorithm quantum parallel computing process according to an embodiment of the present invention.

FIG. 5 is a schematic diagram of the RNUQS algorithm of an embodiment of the invention.

FIG. 6 shows a sum of G transitions in the RNUQS algorithm according to an embodiment of the present inventionGeometric schematic of sub-G transform.

Detailed Description

In order to facilitate the understanding and implementation of the present invention for those of ordinary skill in the art, the present invention is further described in detail with reference to the accompanying drawings and examples, it is to be understood that the embodiments described herein are merely illustrative and explanatory of the present invention and are not restrictive thereof.

The invention provides a real-time dynamic multi-intersection path navigation quantum search method for an urban traffic network, aiming at effectively relieving traffic jam and providing real-time path navigation for running vehicles. The method is used for calculating the path utility value of a large number of vehicles with multiple intersections in the urban road network, and has the advantages that a large number of factors need to be considered, the factors include objective attributes and subjective preference of a driver on the road, the cost corresponding to route selection needs to be considered, and uncertain factors such as emergency possibly occurring on the road need to be considered. And carrying out real-time calculation and search on the influence factors and the path navigation scheme by using quantum calculation and quantum search to obtain a proper utility value and a corresponding path navigation scheme, thereby realizing maximization of the utilization rate of road resources of the whole urban road network while meeting the individual benefits of running vehicles.

The invention maps a real road network (as shown in figure 1(a)) into a model graph R (B, E) (as shown in figure 1(B)), wherein B is a node, E is a vector arrow with a direction between nodes, and R is a graph consisting of B and E. Intersections in FIG. 1(a) are sequentially mapped to node Bs in FIG. 1(B)₁,B₂,...,B₁₂The road links in fig. 1(a) are mapped as vector arrows with directions in fig. 1(b), and the real road network in fig. 1(a) is mapped as a graph R in fig. 1 (b). Each node B represents an intersection in FIG. 1(a), and the node B_i(i 1, 2.., r) denotes the ith intersection, where r is the total number of intersections and each vector arrow E denotes a road segment. Suppose there are n vehicles in the road network, and any vehicle w has a current starting point P_sAnd destination endpoint P_dThen a feasible path of the vehicle can be represented by a continuous adjacent intersection as { P }_s,...,P_i,...,P_d}. Each vehicle selects a feasible path, and the driving paths of all vehicles form a feasible path set FPS_nI.e. a path navigation scheme. The number of vehicles and the number of feasible paths of each vehicle are large, so that the number of path navigation schemes is large, and the problems to be solved by the invention can be converted into searching the optimal path, namely solving the optimal FPS_n. In a real road network, because road conditions are constantly changed, the searching process of the path navigation scheme must be updated in real time within a certain time period to ensure the validity of the path navigation scheme, and therefore, the optimal path must be searched and updated in real time within a limited time.

The utility value U is used for evaluating the advantages and disadvantages of the path navigation scheme, and factors influencing the magnitude of the utility value are many, including unchangeable factors such as the number of lanes, the speed limit of the road sections, the time length of traffic lights, the compliance degree of a driver to the recommended navigation scheme and the like, and also including factors which change continuously along with time such as the distance of a selectable path, time consumption, road conditions and the like, the unchangeable factors are integrated into a preference value P, the changed factors are integrated into a cost value C, and the calculation formula of the utility value is shown as a formula (1).

U＝P-C (1)

The utility value U is an important index for evaluating the quality of the path navigation scheme, namely when the path navigation scheme is determined, the U value is also determined, and the larger the U value is, the better the path navigation scheme is. As known from equation (1), the U value depends on the preference value P and the cost value C, the influence factor of the preference value is shown in table 1, the influence factor of the cost value is shown in table 2, and the U value is determined by the factors in tables 1 and 2 when the path navigation scheme is determined.

TABLE 1 definition of influencing factors and parameters of preference values P

TABLE 2 influence factor and parameter definition of cost value C

The influence degrees of various factors in table 1 and table 2 on the utility value U are different, so in the process of calculating the U value, each factor is given a corresponding weight according to the city scale and the route navigation target. The preference value P is a deterministic factor, and table 1 defines factors that affect the preference value, which is determined after any vehicle has determined a starting location and a destination. Thus, the formula for calculating the preference value P of a certain link is shown in equation (2).

P＝Fr×(α1×Rs+α2×Sl+α3×Ls+α4×Os+α5×Fd) (2)

wherein alpha is_iEach of (i ═ 1, 2., 5) is an independent multiplication factor corresponding to each influence factor, and the value of each multiplication factor is related to the city scale and the setting of a decision target, and the larger the value of the multiplication factor is, the more important the factor is, the larger the influence on the utility value U is, and all the factor values are determined in the same traffic network. In any optional path, the preference value accumulation of each road section is the preference value of the path, and the greater the preference value is, the more optimal the path is.

The magnitude of the preference value P is determined for each road, the magnitude of the cost value C is not only related to the selected route itself, but also takes into account the interplay between vehicles, and table 2 defines the factors of influence of the cost values. After a vehicle determines a route, the values of Ta, De and Tl can be calculated accordingly, but the oil cost Oc is comprehensively determined by the time cost Tc, the distance cost De and the driving speed, but the value of the time cost Tc is not easy to obtain and calculate, because the time spent is not only influenced by the length of the route but also the congestion degree of each road section in the route, the number of vehicles on the road section has a direct relation with the congestion coefficient of the road, and the congestion coefficient of the road is inversely related with the average driving speed of the road. For a specific road section, the average running speed of the road section can be estimated through the number of running vehicles, the traffic congestion coefficient gamma is used for representing the congestion condition of the road, the average running speed of the vehicles on the road is closely related to the traffic congestion coefficient, the actual number of vehicles on the road section is n, the threshold capacity is H, the congestion capacity is L, and the congestion coefficient gamma is calculated as shown in the formula (3).

After the time cost Tc is determined, the oil mass cost Oc and the cost value C may also be calculated. When the feasible paths of all vehicles are determined, the cost value C of any one path can be calculated according to the formula (4).

C＝β₁×Ta+β₂×Tc+β₃×De+β₄×Oc+β₅×Tl (4)

wherein, beta_i(i 1, 2., 5) is an independent multiplication factor of each influencing factor influencing the cost value C, the value of which is related to the city size and the decision target setting, and the size of which represents the degree of influence of each influencing factor on the cost value C and the importance degree thereof. When the cost value C of a certain route is smaller, the route is better, and the sum of the cost values of all vehicles is the final cost value of the route navigation scheme.

The utility value U can measure the quality of a path navigation scheme, and the high utility value is also an important characteristic of good operation of a traffic system. When the driving paths of all vehicles (namely a path navigation scheme) are determined, the utility values can be calculated, the average utility value of the vehicles represents the advantages and disadvantages of the path navigation scheme of the vehicles, and the higher the utility value is, the better the navigation scheme is and the better the traffic condition is. When the utility values U of all possible navigation schemes are obtained, the optimal utility value U is selected_maxThe route navigation scheme performs vehicle guidance, and realizes optimal traffic navigation. However, for a large urban road network, the number of all possible path navigation solutions is huge, and when a common computer is used for calculation and search, due to the limitations of calculation speed and search speed, the real-time performance of vehicle scheduling of the whole road network cannot be realized. The best navigation scheme searched in a limited time has practical application value. Therefore, the invention provides an RGQS (real time dynamic multi-intersection path navigation quantum search method) for an urban traffic network, as shown in fig. 2, the flow of the RGQS method is shown in table 3; the RGQS method consists of UVCQC algorithm and RUNQS algorithm. The parallel capability of the quantum computer and the search capability of the quantum search algorithm break through the limits of the calculation speed and the search speed, and the real-time performance of the whole road network path navigation is realized.

TABLE 3 RGQS method protocol

If there are n vehicles in the road network, the serial numbers are V respectively₁,V₂,...,V_nAny vehicle V_iEach of the n vehicles has its own starting point and ending point, one or more optional paths between the starting point and the ending point are all selected by the system for the driver to meet the requirements of the driver, and the number of the optional paths of the n vehicles is respectively set as b₁,b₂,...,b_nThe path is represented by a crossing set, V_i,jRepresenting the jth path of the ith vehicle, and a path navigation scheme is to extract a set of paths for each vehicle, such as a set(wherein, a)_i(i 1, 2.., n) represents any optional path in the ith vehicle) is a path navigation scheme.

The path navigation scheme corresponds to the utility values thereof one by one, the path navigation scheme is an independent variable x, the utility value U is a function, and the functional relationship is expressed as shown in formula (5).

U(x)＝P(x)-C(x) (5)；

Wherein the value range of the independent variable x isThe form of x in the computer is represented by bits 0 and 1, each value of x uniquely represents a path navigation scheme, for clearer representation, the binary bits of the path navigation scheme x need to represent the path selected by each vehicle, and then the selectable paths of each vehicle need to be represented by a certain number of binary bits, and the formula max { b } is satisfied₁,b₂,...,b_n}≤2^hThe minimum h value is used for coding x, the number of paths of each vehicle needs h-bit binary representation, the total number of paths navigation schemes needs n multiplied by h-bit binary representation, if the number of vehicles is 1000, the paths of each vehicle are coded by 3 bits, 3000 bits are needed to represent one path navigation scheme, and the space needed for storing the schemes is 2³⁰⁰⁰Each ratioParticularly, the method is used for preparing the high-performance liquid crystal display. Classic computers cannot store and even calculate. While quantum computers have excellent performance in data storage and parallel operation, due to the existence of superposition state, 3000 quantum bits can store data which is theoretically 2³⁰⁰⁰Compared with a classical computer, the quantum computer has almost no upper limit on the storage capacity, so that the storage problem of the path navigation scheme can be solved, and the best performance of the quantum computer lies in parallel computation (all independent variables are operated simultaneously, and all function values are obtained by running once) in a true sense on continuous variables. Parallel computation of utility values can therefore be solved.

The basic units stored in the quantum computer are states, only 0 and 1 in the classical computer, and the superposition state exists in the quantum computer, namely the superposition state which can be neither 0 nor 1 exists, so that a 3000-bit binary system in the classical computer can only represent a path navigation scheme x_iAnd 3000 qubits can represent 2 in a quantum computer³⁰⁰⁰Such a path navigation scheme, as long as the quantum state is not observed, may consider these 2 s³⁰⁰⁰The path navigation schemes are stored simultaneously, the quantum computer is very suitable for storing the continuous variables, and each path navigation scheme exists in the same probability which is expressed by a probability amplitude sigma in quantum mechanics, and the square sigma of the probability amplitude of a certain path navigation scheme²Equal to the probability that the path navigation solution can be output (observed at the output).

In the invention, an independent variable x is used for representing a path navigation scheme, a function U (x) is used for representing a utility value of the navigation scheme, in a quantum computer, the function U (x) and the function U (x) are both represented by quantum states and are respectively stored by two registers, and the independent variable x is initialized as shown in a formula (6).

As shown in FIG. 3, the route guidance scheme is determined by the number of vehicles n and the start and end point information of the vehicle, and can be twoThe binary code represents all path navigation schemes, and the total number of the navigation schemes is less than or equal to 2^n×hLet S be 2^n×hTherefore, all path navigation schemes can be completely represented by S quantum states. The equal weight superposition state of the argument x (i.e., all path navigation schemes) is the input to the quantum computation, equation (6)The probability amplitude sigma (its square sigma) representing the existence of the path navigation solution²Representing the probability of the corresponding path navigation scheme), the S states respectively represent the values of the N path navigation schemes x, wherein S is equal to N, and the existence probabilities of all the path navigation schemes in the superposition state are allThe quantum state in formula (6) is in shorthand form, such as state |0>Is in fact thatThe total digits of all states are n x h digits, n represents the number of vehicles, the selectable path of each vehicle is represented by h-digit binary, if the ith h digit is all 0, the selected path of the ith vehicle is the first path (numbered 0), the S quantity states comprehensively represent all path navigation schemes, and the storage and the input of the path navigation schemes can be effectively solved. The calculation of the function U (x) in the UVCQC algorithm is shown in equation (7).

Different quantum circuits are needed for different quantum calculations, the quantum circuits need to be determined according to functions, and a unitary transform U must be used for calculating the functions in a quantum computer_fThe subscript f refers to a function, different unitary transformations use different quantum lines, and an auxiliary qubit | z is required in a quantum computer>To realize unitary transformation and obtain a function, the specific calculation process is shown as formula (8).

In this transformation, the input is unique for a particular output.

As shown in fig. 3, after the utility value function U (x) is determined, appropriate quantum lines and auxiliary quantum bits need to be set according to U (x) to implement unitary transformation, each state, that is, the independent variable x, corresponds to one utility function value U (x), all independent variables simultaneously perform the same operation, and the utility value calculation is completed in parallel, and the calculation process can be obtained from the quantum mechanics property as shown in formula (9).

The result of running once in a particular line for quantum computation is shown in equation (10). Fig. 4 is a process of quantum parallel computation of the UVCQC algorithm once.

All U values are stored in another register, and assuming i > is observed at one end of the argument, the register storing the U value is also collapsed to | U (i) >, and after i is observed, the probability that the value observed by the register storing the U value is U (i) is 1, and vice versa. Through the above analysis, a UVCQC algorithm flow can be obtained as shown in table 4.

TABLE 4 UVCQC Algorithm

The above-mentioned solution is to obtain and calculate each influence factor of the utility value U, however, although the quantum computer can perform parallel calculation, the extraction of the result is not easy, and the result must be single-output, that is, the quantum computer can calculate the utility values of all navigation solutions, but the utility values will collapse when observed at the output end, and only one output utility value can be obtained finally. For the urban road network path navigation problem, only an optimal path navigation scheme needs to be obtained, and therefore only one corresponding utility value (optimal utility value) needs to be obtained. The utility value of massive disorder is obtained, and the quantum computer searches massive disordered data by an efficient algorithm, namely a quantum search algorithm. However, the quantum search algorithm can only solve the situation that a target state (the target state refers to a state to be searched, and the state corresponding to the optimal utility value is referred to herein) is determined, and cannot be successfully searched by one hundred percent.

The path navigation problem is converted into an optimal result searching problem in the equal weight superposition state for obtaining the utility value, and the searching set is { | U>}＝{|U(0)>,|U(1)>,...,|U(N-1)>The number of the utility value states is S, and the target state (i.e. the state needing to be output) is U_max(maximum utility value), the target state is unknown, so that the maximum utility value and the corresponding path navigation scheme cannot be obtained directly through the quantum search algorithm. In a real road network, the minimum utility value which does not cause congestion may be regarded as a fixed empirical value k, and then utility values larger than the empirical value k may be output as a result, and if the number of utility values larger than the empirical value k is m, then any one of m satisfies the output condition.

The number of target states is m, and the function used by the RNUQS algorithm to determine the target state is called oracle function, where y is U (x), and the oracle function used by RNUQS is shown in equation (11).

After the utility value function state | U (x) > is distinguished by the oracle function, the state with the function value f (x) of 1 is the target state, m target states are locked accordingly, the RNUQS algorithm judges whether the state is the target state according to whether the function value corresponding to the state is 1, and the RNUQS algorithm can obtain correct output by improving the probability amplitude of the target state.

In the searching process, an oracle function is used for checking whether each utility value is in a target state, then the probability range of the target state is enlarged through Grover transformation to improve the probability of the output of the target utility value state, as shown in FIG. 5, G in the graph represents Grover transformation, G transformation is abbreviated in the following explanation, G transformation is carried out once, namely, specific quantum iteration is carried out once, after G transformation of a certain number of generations, the probability of the target utility value state is increased to a certain degree, and finally, the target utility value state is output with the probability close to 1, so that the proper target utility value state is obtained.

Where G ═ DO, O denotes an oracle query, let us | U_a>Is an object state that, upon query with an oracle, will perform the unitary transformation I-2| U_a><U_aIf the target state is not the target state, this operation is not performed, so the calculation of O is shown in equation (12).

O＝I-2|U_a><U_a| (12)

The calculation of D is shown in equation (13).

WhereinIs an equally weighted stack of all the basic states,h denotes the Hadamard transform (implemented with Hadamard gates), used to prepare the equal-weight superposition states,representing the preparation of an equal weight superposition state of n x h bits. After the initial equal-weight superposition state is subjected to G transformation every time, the probability amplitude of the target utility value state is increased by one point, the probability amplitude of the non-target utility value state is decreased by one point, after the G transformation of certain iteration times, the output probability of the target utility value state is close to 1, and at the moment, the target utility value state can be observed at the output end to obtain a proper utility value.

In order to better understand the role of one G transformation, one G transformation can be regarded as quantum transformation of quantum states in a two-dimensional space and is divided into two steps, namely O transformation and D transformation. FIG. 6(a) is a geometric diagram of a G-transform, and FIG. 6(b) is a diagram of a G-transformGeometric schematic of sub-G transform, | U_a>Is the target state, the projection of the current superposition state on the target state represents the output probability amplitude of the target state in the superposition state, and the original state rotates to the target state by an angle of 2 theta every time G transformation is carried out, as shown in figure 6(a),the process of the sub-G transformation is shown in FIG. 6(b), where the angle α in FIG. 6(a) is an arbitrary acute angle, and the angles θ in FIGS. 6(a) and 6(b) are equal (b))。

In the context of figure 6(a),is an initial equal weight superposition state, | U_t>Representing any current state, all G transitions are for | U_t>Change, | U_a>Represents the sum of all target states, calculated as equation (14)Shown in the figure.

a_iThe state of the object is represented by,represents | U_a>Is in an orthogonal state with | U_a>Perpendicular, | U_t>Andthe included angle of the angle is set as alpha,andis theta, equal weight superposition stateIn target state | U_a>The projection (probability amplitude) of The meaning is that the probability of observing the target state under the equal weight superposition state is sin²θ is m/N, and the current state is | U_t>After one G conversion, the current state is converted into O | U_t>，|U_t>And O | U_t>AboutSymmetrical, O | U_t>Then transformed into G | U through D transformation_t>，O|U_t>And G | U_t>AboutSymmetry according to the angleCalculation, G | U_t>And | U_t>the included angle of the angle is 2 theta, which is irrelevant to alpha, and the current state rotates counterclockwise by the angle 2 theta every time G transformation is carried out.

Due to the utility value state | U>Initially in an equal weight superposition state, after i times of G transformations, andbecomes (2i +1) theta, and in order to make the target state output with a probability close to 1, (2i +1) theta is made to be approximately equal to 1, wherein Is calculated toround represents the nearest integer, so that a proper target utility value can be searched only by performing i times of transformation, and the required time complexity is only

It can be seen from the calculation of the value of i that since i can only take integers, the probability that the target state can be finally obtained is only very close to 1, so that there is a possibility that an output error occurs, and in the actual path navigation, the error is not allowed. To address this problem, the present invention proposes a Quantum Error Detection Strategy (QEDS), and the QEDS Strategy flow is shown in table 5. The empirical value k can only ensure that a proper output is provided, but the output cannot ensure enough optimization, and an ideal empirical value k can be set in practical situations_mPerforming multiple searches, and meeting the maximum time limit t of real-time performance_maxOn the premise of (1), search is carried out as many times as possible, and the time spent for one search is set as t_cCurrently it has taken time t_s(initially 0).

TABLE 5 QEDS policy

Thus, the RNUQS algorithm flow proposed by the present invention is shown in table 6.

TABLE 6 RNUQS Algorithm

It should be understood that parts of the specification not set forth in detail are well within the prior art.

It should be understood that the above description of the preferred embodiments is given for clarity and not for any purpose of limitation, and that various changes, substitutions and alterations can be made herein without departing from the spirit and scope of the invention as defined by the appended claims.

Claims

1. A real road network real-time dynamic multi-intersection path navigation quantum search method maps a real road network into a model graph R (B, E), wherein B represents an intersection node set, and B represents an intersection node set_iRepresents a single intersection node, i ═ 1, 2.., r; r is the total number of road junctions, and E represents a road section set with directions; suppose there are n vehicles in the road network, and any vehicle w has a current starting point P_sAnd destination endpoint P_dThen a feasible path of the vehicle is represented by the nodes of the continuous adjacent intersections as { P_s,...,P_i,...,P_d}; each vehicle selects a feasible path, soThe driving path with vehicle forms a feasible path set FPS_nI.e. a path navigation scheme;

characterized in that the method comprises the following steps:

step 1: initializing a vehicle set { v) according to the number n of vehicles, start and stop point information and an optional path of each vehicle₁,v₂,...,v_nAnd optional Path setWherein v is_iIt means that the (i) th vehicle,one of all alternative paths representing the ith vehicle;

step 2: for vehicles and their alternative paths 0,1_iCarry out quantum coding { |0>,|1>,...,|2^n×h-1>Determining that the quantum state can completely represent all path navigation schemes; wherein b is_iRepresenting the number of selectable paths of the ith vehicle, and h representing the minimum number of binary digits required for encoding the selectable paths;

2. The method for real-time dynamic multi-intersection path navigation quantum search of the urban traffic network according to claim 1, wherein the utility value function U (x) in step 3 is:

wherein Fr (x) represents whether the road section can be reached, 1 represents reachable, and 0 represents unreachable; rs (x) represents the road section condition and takes the value of [0, 1%](ii) a Sl (x) represents a speed limit and takes the value [0, 1%](ii) a Ls (x) represents the illumination condition of the road section and takes the value of [0, 1%](ii) a Os (x) represents driver compliance with system recommendations, and takes the value [0, 1%](ii) a Fd (x) represents the familiarity of the driver with the road section and takes the value of [0, 1%](ii) a Ta (x) represents the road influence caused by sudden traffic accidents or temporary control, and takes the value of [0, 1%](ii) a Tc (x) represents the time cost spent by the selected path, taking the values [0, ∞](ii) a De (x) represents the distance cost consumed by the selected path and takes the values of [0, ∞](ii) a And Oc (x) represents the oil consumption cost of the selected path and takes the values of [0, ∞ ]](ii) a Tl (x) represents the influence of traffic lights and takes the value [0, 1%]；α_i、β_iEach of the factors represents an independent multiplication factor, i is 1, 2.

3. The urban traffic network real-time dynamic multi-intersection path navigation quantum search method according to claim 1, wherein the specific implementation of step 4 comprises the following substeps:

4. The urban traffic network real-time dynamic multi-intersection path navigation quantum search method according to claim 1, wherein the specific implementation of step 5 comprises the following substeps:

Wherein, a_iRepresents the target state, | a_i>Quantum form representing the ith target state, m representing the total number of target states;

O＝I-2|U_a><U_a|；

5. The urban traffic network real-time dynamic multi-intersection path navigation quantum search method according to claim 4, wherein the specific implementation of step 5.7 comprises the following substeps:

Step 5.7.2: if t is_s＜t_maxThen the following step 5.7.3 is performed, where t_maxRepresenting a maximum navigation time interval that can guarantee the real-time performance of path navigation; otherwise, execute downStep 5.7.5;

step 5.7.3: if U is present_out< k, then t_s＝t_s+t_cAnd go back to perform said step 5.7.2, where t_cRepresenting the time required for executing the urban traffic network real-time dynamic multi-intersection path navigation quantum search method once; otherwise, the following step 5.7.4 is performed;

step 5.7.4: fruit of Luo Guo U_out＜k_mThen k is equal to U_out，t_s＝t_s+t_cAnd go back to perform said step 5.7.2, where k_mIndicating an ideal utility value set empirically; otherwise, the following step 5.7.5 is performed;

step 5.7.5: u shape_s＝U_outOutput U_s。