Wireless ad hoc network routing and resource cross-layer joint optimization method based on genetic algorithm

Abstract

The invention discloses a wireless ad hoc network routing and resource cross-layer joint optimization method based on a genetic algorithm. The method comprises the following steps: step 1, aiming at a plurality of end-to-end communication services in a wireless ad hoc network, carrying out mathematical modeling on a normalized difference sum of each service flow path hop count and a theoretical minimum value thereof and a constraint condition; performing joint optimization on link layer resource allocation and network layer routing decision; and step 2, solving the mathematical model based on a genetic algorithm to realize the targets of reasonable distribution of network resources and optimization of service quality. According to the method, link layer resource allocation and network layer routing decision are jointly optimized in a global mode under the condition that network communication resources are limited, it is ensured that a service flow is close to shortest path transmission as far as possible, end-to-end delay and network loads are reduced, and effective allocation of resources and flexible construction of routing are achieved.

Description

Technical Field

[0001] The present invention belongs to the technical field of communication network protocol layering, and particularly relates to a cross-layer joint optimization method for wireless ad hoc network routing and resources based on genetic algorithms. Background Art

[0002] In the traditional communication network protocol layering system, the constructed protocol stack is often based on the upward support of the lower protocol layer for the higher protocol layer. When the network layer manages communication routing, it needs to further optimize the routing decision and allocation of this layer based on the specific state of the network link that has been completed by the data link layer. This layering method ignores the coordination and cooperation between layers and is not conducive to the overall network resource optimization. In some scenarios where the communication resources of nodes and links are limited, the method of scheduling node and link resources according to the routing decision result is not applicable.

[0003] A wireless ad hoc network is a distributed, multi-hop autonomous network system composed of a group of mobile terminals or fixed terminals with both transmitting and receiving functions, which is different from traditional single-hop wireless networks. The communication within the system is achieved through the cooperative forwarding of each node without the support of fixed infrastructure. Each node in the network is both a communication terminal for transmitting and receiving and a router for assisting in the routing process. This network system has strong autonomy and flexibility and is easy to deploy. Wireless ad hoc networks have shown great potential and advantages in many special application fields, such as military communication, disaster relief, intelligent transportation systems, etc. Compared with traditional fixed infrastructure networks, the network environment of MANET is more complex. The high mobility of nodes, limited wireless resources, dynamically changing network topology, and the characteristics of multi-hop communication make it face many challenges. Especially in the case where communication resources such as nodes and links are limited, how to optimize the network routing decision, improve communication quality, reduce network load, and ensure the stability and reliability of the overall network has become the core issue of current research.

[0004] Yan Wei of Hangzhou Dianzi University proposed a distributed Newton algorithm for utility optimization in an AdHoc network in the paper on cross-layer optimization design based on the AdHoc network. The energy consumption cost was introduced into the utility maximization function, and cross-layer design was carried out by jointly controlling the rate and power in the network layer, data link layer, and physical layer. Its main defects are as follows: The Newton algorithm assumes that the network state (topology, channel) is constant within the iteration period. However, due to the node mobility and channel time-variability in the AdHoc network, the optimization model parameters drift frequently. When the topology change frequency exceeds the algorithm convergence speed, the fluctuation amplitude of the objective function of the Newton method is relatively large; The convergence of the Newton method highly depends on the initial point selection. In the AdHoc network, random initialization may cause the algorithm to fall into a "bad" local extreme value, such as over-preferring short-hop paths and ignoring slot conflicts.

[0005] In the existing hierarchical AdHoc network design, protocol layers work relatively independently, and adjacent layers are connected through static interfaces. The design of each layer is based on the implementation of lower-layer functions to further implement additional functions. The advantage of this design pattern is good robustness and simple design, but there are also some deficiencies. For example, the scheduling and allocation of link resources in the link layer will affect the decision-making of the network layer, and the path decision-making and routing selection in the network layer will also affect the resource allocation in the link layer, thus affecting each other. Summary of the Invention

[0006] To overcome the above-mentioned deficiencies in the prior art, the purpose of the present invention is to provide a cross-layer joint optimization method for routing and resources in a wireless ad hoc network based on a genetic algorithm. In a global manner, under the condition of limited network communication resources (node time slot resources), the link layer resource allocation and network layer routing decision-making are jointly optimized to ensure that the service flow is as close as possible to the shortest path transmission, reduce the end-to-end delay and network load, and achieve effective resource allocation and flexible route construction.

[0007] To achieve the above purpose, the technical solution adopted by the present invention is:

[0008] A cross-layer joint optimization method for routing and resources in a wireless ad hoc network based on a genetic algorithm, including the following steps;

[0009] Step 1: For several end-to-end communication services in the wireless ad hoc network, a mathematical model is established for the sum of the normalized differences between the number of hops of each service flow path and its theoretical minimum value, as well as the constraint conditions (flow conservation constraint, node time slot resource capacity constraint). This is used to solve the problem of jointly optimizing the link layer resource allocation and network layer routing decision-making in the wireless ad hoc network under the scenario of limited node time slot resources, ensuring that the service flow is as close as possible to the shortest path transmission, and reducing the end-to-end delay and network load;

[0010] Step 2: Solve the model established in Step 1 based on the genetic algorithm to achieve the goals of network resource allocation and service quality optimization.

[0011] The specific content of Step 1 is as follows:

[0012] Step 1.1: Specifically define the variables and metrics in the network;

[0013] Suppose there are k network nodes with the same function in the network, each frame period has the same time slot resource C, and the number of network services is N;

[0014] Define a k×k network connectivity matrix A, where each element a ij (i,j∈[1,k]) represents whether there is a link from node i to node j in the network, and b ijis 1, representing the existence of a link, and is 0 representing the non - existence of a link. For an undirected network, there is a ij = a ji ;

[0015] Define the k×k matrix B as the link selection matrix of a certain service path, where the element b ij represents whether the link from node i to node j is selected in this end - to - end path of the service. If it is 1, it means this link is selected; if it is 0, it means this link is not selected. For an undirected network, there is b ij = b ji ;

[0016] Define the k×k matrix C as the effective link selection matrix of a certain service path, where the element c ij is the selected link and ensures that this link is connected and exists. Therefore, it can be seen that c ij = a ij ×b ij , and at the same time, if c ij = 1, then this link is a routing component of this service, and this link is allocated a unit of time - slot resource for the corresponding service. For an undirected network, there is c ij = c ji ;

[0017] Define the normalized cost function of the end - to - end communication service m (m∈[1,N]) with the same load, where F (m) is:

[0018]

[0019] Among them, the parameter H (m) represents the number of hops of the decision path for service m, and the parameter represents the number of hops of the shortest path for service m;

[0020]

[0021] Among them, the parameter k is the number of network nodes, is the element in the path effective link selection matrix of service m;

[0022] Step 1.2: Model the routing decision problem of resource optimal allocation of multiple services in the network;

[0023] Under the condition of limited time - slot resources, based on the minimum total cost of the selection paths of N services by the network, formulate the performance objective function as:

[0024]

[0025] To achieve this performance goal, it is necessary to further set its constraints. The constraints of this objective optimization problem are path connectivity constraints and node time-slot resource constraints.

[0026] In step 1.1, matrix A describes the physical connection status of the network, providing underlying constraints for routing optimization. Matrix B encodes the end-to-end path selection strategy of traffic flows, independent of the actual network state. Matrix C synthesizes physical connectivity and traffic demands to screen out the actually available routing links. The relationship between them is as follows:

[0027] C = A ⊙ B (element-wise multiplication)

[0028] Only when a ij = 1 and b ij = 1, c ij = 1, otherwise it is 0, indicating that the validity of the service path needs to satisfy both physical connectivity and routing decisions simultaneously.

[0029] In step 1.2:

[0030] The path connectivity constraint is:

[0031]

[0032] where s and d are the source node and sink node of a certain service respectively, and i, j ∈ [1, k];

[0033] To ensure the connectivity of the path selected by the service, for the source node and sink node in the path selected by a certain service, there is exactly one other node in this path that is connected to it through a link. For the other nodes passed by in the path selected by the service, there must be two nodes in this path that are connected to it.

[0034] In step 1.2:

[0035] The node time-slot resource constraint is:

[0036]

[0037] where represents the time-slot resource usage of the paths selected by N services for any node j.

[0038] To solve the network cross-layer routing problem of joint routing and time-slot resource allocation, a genetic algorithm is introduced here. The number of hops of the service path and the total node resource constraint are incorporated into the genetic algorithm framework, and a penalty function is combined to avoid resource overrun. The optimal solution is generated through crossover and mutation operations, significantly reducing the cost of the network serving the service group.

[0039] Step 2 is specifically as follows:

[0040] Step 2.1: Generate a set of candidate paths, and calculate the shortest path for each service using the Dijkstra algorithm and generate the top n shortest paths through Yen's algorithm; this algorithm generates a set of candidate paths for each service m for selection by the genetic algorithm;

[0041] The algorithm steps of the said Step 2.1 are as follows:

[0042] ① Initialization:

[0043] Use the Dijkstra algorithm to calculate the first shortest path P1;

[0044] Initialize the candidate path list P = {P1} and the deviation node queue D;

[0045] ② Iteratively generate paths (for the i-th path, i ∈ [1, x - 1]):

[0046] Select the current path: Take out the path P from P i ;

[0047] Path deviation:

[0048] For each node v of path P i (except the destination node): j (except the destination node):

[0049] ◆ Root path: Intercept P i from the source node to the sub-path Root = s → v1 →... → v of v j ; j ;

[0050] ◆ Forbidden edge: Forbid using the edge (v i , v j ), v j , v j +1) starting from v in P;

[0051] ◆ Deviation path calculation: Starting from v j , use the Dijkstra algorithm to find the shortest path Spur to the destination node d in the remaining graph;

[0052] ◆ Merge path: If Root + Spur is valid and does not exist in the candidate list, add it to the candidate queue D;

[0053] Select the next shortest path: Select the shortest path from D and add it to P;

[0054] ③ Termination condition:

[0055] Terminate when k paths are found or the candidate queue D is empty;

[0056] At this time, generate the set P of the first n shortest paths for service m (m) , ensuring the diversity of the genetic algorithm in path selection.

[0057] Step 2.2: Define the chromosome coding rule. A chromosome consists of N genes, each gene corresponding to a service, with the value range [1, n], and the gene value g m = x indicates that service m selects path

[0058] For example: For N = 3 services and n = 5, the chromosome Chr = [2, 5, 1] means:

[0059] Service 1 selects the 2nd candidate path

[0060] Service 2 selects the 5th candidate path

[0061] Service 3 selects the 1st candidate path

[0062] Step 2.3: Design the fitness function. The original problem is to minimize F, while the genetic algorithm usually maximizes the fitness. Therefore, F needs to be converted into a positive index:

[0063] Step 2.4: Initialize the population, generating feasible chromosomes through random generation and constraint correction;

[0064] Step 2.5: Iteratively perform selection, crossover, and mutation operations, and update the population in combination with the elite retention strategy;

[0065] ① Selection operation (tournament selection)

[0066] ◆ Randomly select t individuals (usually t = 3) to form a tournament group;

[0067] ◆ Select the individual with the highest fitness in the group to enter the next generation;

[0068] This operation can maintain the population diversity and avoid premature convergence;

[0069] ② Crossover operation (single-point crossover)

[0070] ◆ Randomly select two parent chromosomes Chr1, Chr2;

[0071] v Randomly select a crossover point c ∈ [1, N - 1];

[0072] ◆ Generate offspring

[0073] ◆ Perform constraint correction on the offspring (same as the constraint correction in step (2.4));

[0074] ③ Mutation operation (path replacement)

[0075] ◆ For each gene g m , with probability p m (usually p m = 0.1) trigger mutation;

[0076] ◆ Randomly select a new path number g′ m ∈ [1, n], and g′ m ≠ g m ;

[0077] ◆ Correct the constraints of the individual after mutation;

[0078] ④ Elite retention strategy

[0079] Retain the top e individuals with fitness in each generation (usually e = 5% × population size) and directly pass them to the next generation to prevent the loss of excellent genes.

[0080] Step 2.6: Set the maximum number of iterations, output the optimal chromosome, and parse it into the routing path and time slot allocation scheme for each service.

[0081] The specific content of step 2.3 is as follows:

[0082]

[0083] At the same time, resource constraints need to be handled through penalty terms, and a penalty function is defined:

[0084]

[0085] If the resource occupancy of a node exceeds the capacity C, then accumulate the exceeded part (max(0, ·) to ensure non - negativity).

[0086] Finally, introduce a weight coefficient β to balance the importance of the objective and constraints, and its expression is:

[0087]

[0088] The specific content of step 2.4 is as follows:

[0089] First, perform random generation. For each chromosome, randomly assign a path number g m ∈ [1, n] for service m. Then perform constraint correction. If the load of node i

[0090]

[0091] Adjust according to the following priorities:

[0092] Replace the path of the service with the highest load, and select a candidate path that does not contain node i. If there is no selectable path, randomly disconnect an associated link of node i to force path detour. At the same time, use depth-first search (DFS) to detect path breaks and replace the broken link with a valid link in the candidate path.

[0093] For example, if initializing the chromosome causes the load of node 5, U(5) = 12 > C = 10, then select the service passing through node 5 and replace its path with a sub-optimal path that does not pass through this node.

[0094] Step 2.6 is specifically as follows:

[0095] Map the gene values to the paths of each service {P

[0096] , P (1) , P (2) , …, P (N)}. Since for all c ij = 1, it means that this link is a routing component of this service, and this link allocates a unit of time slot resource for the corresponding service. Therefore, according to the generated path set, the time slot allocation scheme of each node can be obtained.

[0097] The method is used for intelligent transportation and vehicular ad hoc networks (VANETs), disaster emergency communication, unmanned aerial vehicle networks (UAVs), Internet of Things (IoT), and mobile ad hoc networks (MANETs).

[0098] Advantages of the present invention:

[0099] First, joint optimization can comprehensively consider the resource requirements of the network layer and the link layer. For example, in the case of an unmanned aerial vehicle emergency communication network, the requirements of this scenario are rapid deployment, low latency, and high reliability. While ensuring a reduction in end-to-end latency, this invention can avoid resource conflicts or imbalances caused by separate optimization, thereby reducing the problem of overloaded node loads and improving the stability and performance of the network.

[0100] Second, through the crossover and mutation operations of the genetic algorithm, the search space is expanded to explore better solutions. Combining the elitist retention strategy ensures that high-quality solutions are retained during iteration, effectively balancing path hop count optimization and resource conflict suppression, and realizing the joint optimization of multi-service routing decision-making and resource allocation. Brief Description of the Drawings

[0101] Figure 1 It is a flowchart of the method of the present invention. Detailed Embodiment

[0102] The present invention will be further described in detail below with reference to the accompanying drawings.

[0103] Network scenario:

[0104] In a wireless ad hoc network, the node devices are exactly the same, with only one set of transceiver channels, and omnidirectional antennas are used for wireless transmission and reception. Each node uses the same TDMA method to plan time slots and frame periods. Therefore, the time slot resources of nodes and links are limited.

[0105] The node devices use the TDD method to process transmission and reception information. The transmission and reception time slots are orthogonal in the time dimension and both require time slot resources. The network supports single-path multi-hop relay forwarding services between end-to-end nodes. There are several end-to-end communication services in the network. Each service exists independently and has the same traffic load, and only occupies one time slot of relevant nodes and links within one frame period.

[0106] By allocating time slot resources and making routing decisions for these services, the goal of reasonable allocation of network resources and optimization of service quality is achieved.

[0107] A cross-layer joint optimization method for routing and resources in a wireless ad hoc network based on genetic algorithm, including the following steps;

[0108] Step 1. Mathematically model the problem:

[0109] 1.1 Specifically define the variables and metrics in the network:

[0110] Suppose there are k nodes with the same function in the network. Each node has only one set of transceivers. The node duplex mode is TDD, and the transmission and reception time slots are orthogonal. The network nodes use the TDMA mechanism for time slot management. Each frame period has the same time slot resource C, that is, the sum of the transmission and reception time slot resources of all associated links of this node is limited. The network service is end-to-end single-path multi-hop service, and the number of services is N.

[0111] Define a k×k network connectivity matrix A, where each element a ij (i,j∈[1,k]) represents whether there is a link from node i to node j in the network. If a ij is 1, it means the link exists; if it is 0, it means the link does not exist. For an undirected network, there is a ij =a ji .

[0112] Define a k×k matrix B as the matrix of link selection for a certain service path, where the element b ij represents whether the link from node i to node j is selected in this end-to-end path of the service. If it is 1, it means this link is selected; if it is 0, it means this link is not selected. For an undirected network, there is b ij =b ji .

[0113] Define the k×k matrix C as the effective link selection matrix for a certain service path, where the element c ij is the selected link, and it is ensured that this link is connected and exists. Therefore, it can be seen that c ij = a ij ×b ij . At the same time, if c ij = 1, it means that this link is the routing component of this service, and this link allocates a unit of time slot resource for the corresponding service. For an undirected network, c ij = c ji .

[0114] Among them, the matrix A describes the physical connection status of the network and provides the underlying constraints for routing optimization. The matrix B encodes the end-to-end path selection strategy of the service flow and is independent of the actual network status. The matrix C synthesizes the physical connectivity and service requirements and filters out the actually available routing links. The relationship between them is:

[0115] C = A⊙B (element-wise multiplication)

[0116] Only when a ij = 1 and b ij = 1, c ij = 1, otherwise it is 0. It represents that the effectiveness of the service path needs to satisfy both physical connectivity and routing decision-making simultaneously.

[0117] Define the normalized cost function of service m (m ∈ [1, N]), where F (m) is:

[0118]

[0119] Among them, the parameter H (m) represents the number of hops of the decision path for service m, and the parameter represents the number of hops of the shortest path for service m.

[0120]

[0121] Among them, the parameter k is the number of network nodes, is the element in the path effective link selection matrix of service m.

[0122] 1.2 Model the routing decision problem of resource optimal allocation of multiple services in the network, map the network design goal to the mathematical objective function F, and make the optimization process measurable and comparable. It is used to solve the problem of ensuring that the service flow is as close as possible to the shortest path transmission and reducing the end-to-end delay and network load in the scenario of limited node time slot resources in a wireless ad hoc network.

[0123] Under the condition of limited slot resources, based on the network, the sum of the path costs of selecting N services is minimized, and the performance objective function is formulated as:

[0124]

[0125] To achieve this performance objective, it is necessary to further set its constraint conditions. First, it is necessary to ensure loop-free transmission of data packets from the source to the destination to avoid path breaks or loops. Second, limit the number of traffic flows that a node serves simultaneously to prevent slot resource overload. Therefore, the constraint conditions for this objective optimization problem are:

[0126] (1) Path connectivity constraint:

[0127]

[0128] Among them, s and d are the source node and sink node of a certain service respectively, and i, j ∈ [1, k].

[0129] To ensure the connectivity of the path selected by the service, for the source node and sink node in the path selected by a certain service, there is exactly one other node in this path that is connected to it through a link. For the other nodes passed by in the path selected by the service, there must be two nodes in this path that are connected to it.

[0130] (2) Node slot resource constraint:

[0131]

[0132] Among them represents the slot resource usage of the paths selected by N services for any node j.

[0133] Step 2: Solve the mathematical model based on the genetic algorithm:

[0134] To solve the network cross-layer routing problem of joint routing and slot resource allocation, a genetic algorithm is introduced here. The number of hops of the service path and the total node resource constraint are incorporated into the genetic algorithm framework, and a penalty function is combined to avoid resource overrun. The optimal solution is generated through crossover and mutation operations, significantly reducing the cost for the network to serve the service group.

[0135] Among them, the objective function guides the fitness calculation to ensure that the optimization direction of the algorithm is consistent with the service objective (minimizing the sum of the normalized differences between the number of hops of the traffic flow and its theoretical minimum value); the flow conservation constraint is rigidly satisfied through the chromosome encoding rule to avoid the generation of invalid solutions; the node capacity constraint is softly processed through a dynamic penalty term to balance feasibility and solution quality, achieving the goal of reasonable allocation of network resources and optimized service quality.

[0136] 2.1 Generate a set of candidate paths. By calculating the shortest path and expanding the sub-optimal paths, construct a set of feasible paths for each service to narrow the search space of the genetic algorithm. Calculate the shortest path for each service using the Dijkstra algorithm and generate the top n shortest paths using Yen's algorithm;

[0137] 2.2: Define the chromosome encoding rule to map the service path selection to the genotype representation of the genetic algorithm. The chromosome gene represents the path number selected by the service, and the gene value range is [1, n];

[0138] 2.3: Design a fitness function to quantify the quality of the chromosome and guide the algorithm to evolve in the direction of low hop count and low resource conflict. Its expression is:

[0139]

[0140] where F is the cost function, Penalty is the penalty term for the total node resource exceeding the limit, and β is the penalty coefficient;

[0141]

[0142] where H (m) is the actual path hop count (i.e., the selected path length) of service m.

[0143] Indicator function:

[0144]

[0145] where P (m) represents the path selected by service m

[0146] C: represents the total time slot resource capacity of each node within a single frame period.

[0147] max(0, ·): Only penalize the part where the resources exceed the limit, and the penalty value is 0 when not exceeding the limit.

[0148] 2.4: Initialize the population and generate feasible chromosomes through random generation and constraint correction;

[0149] 2.5: Iteratively execute the selection, crossover, and mutation operations, and update the population in combination with the elite retention strategy;

[0150] 2.6: Output the optimal chromosome and parse it into the routing path and time slot allocation scheme for each service.

[0151] Application Example

[0152] Scenario Description:

[0153] 50 unmanned aerial vehicles (UAVs) are deployed in a certain earthquake-stricken area to form a temporary emergency communication network. Due to wind or obstacle avoidance movement of the UAVs, each UAV only supports forwarding 5 traffic flows simultaneously (C = 5), and requires the lowest possible latency and packet loss rate.

[0154] Steps of the algorithm implementation:

[0155] 1. Initialization phase

[0156] Input parameters: Network connectivity matrix A: The link state is updated in real time through periodic beacon exchanges between UAVs.

[0157] Traffic demand matrix: 30 end-to-end traffic flows are defined, and the shortest path of each flow is pre-computed by the distributed Dijkstra algorithm.

[0158] Genetic algorithm parameters: Population size = 200, Crossover rate = 0.8, Mutation rate = 0.05, β0 = 1, λ = 0.1.

[0159] Chromosome encoding:

[0160] Path gene: Binary matrix B, dimension 50×50, b ij = 1 indicates that the traffic flow selects the link i→j.

[0161] 2. Genetic optimization process

[0162] Fitness calculation:

[0163] In each generation, individuals filter invalid paths through C = A⊙B, and calculate:

[0164]

[0165] Dynamic penalty: β(t) = 1·e 0.1t , allowing over-limit exploration in the initial stage and enforcing strict capacity constraints in the later stage.

[0166] Genetic operations:

[0167] Constraint-aware crossover: Exchange gene segments only on co-activated links (b ij = 1) to ensure that the offspring satisfy flow conservation.

[0168] Load-sensitive mutation: For nodes with a load rate > 80%, disconnect one of its high-hop links with a probability of 0.3 and randomly activate a low-load link.

[0169] Elite retention: Retain the top 20 individuals with the highest fitness in each generation to maintain high-quality genes.

[0170] Compared with the traditional AODV + static TDMA method, this invention can significantly reduce the average end-to-end latency and the maximum node load rate.

[0171] This example verifies the effectiveness of the genetic algorithm in achieving joint optimization of routing and resources in a dynamic UAV network. Through cross-layer matrix coupling, a dynamic penalty mechanism, and load-sensitive mutation, the delay, reliability, and load balance are significantly improved.

Claims

1. A wireless ad hoc network routing and resource cross-layer joint optimization method based on genetic algorithm, characterized in that: The steps include: Step 1: mathematically model several end-to-end communication services in the wireless ad hoc network, and jointly optimize the link layer resource allocation and network layer routing decision in the wireless ad hoc network; Step 2: Solve the model established in step 1 based on the genetic algorithm to achieve the goals of network resource allocation and service quality optimization.

2. The method for cross-layer joint optimization of wireless ad hoc network routing and resources based on genetic algorithm according to claim 1, characterized in that: The step 1 is specifically as follows: Step 1.1: Define the variables and indicators in the network; Assume that there are k network nodes with the same functions in the network, each frame period has the same time slot resources C, and the number of network services is N; Define a k×k network connectivity matrix A, where each element a ij , i,j∈[1,k], represents whether there is a link from node i to node j in the network, a ij is 1, indicating that the link exists, and 0 indicates that the link does not exist. For an undirected network, there is a ij =a ji ; Define the k×k matrix B as the link selection matrix of a certain service path, where the element b ij Indicates whether the link from node i to node j is selected in the end-to-end path of the service. If it is 1, it means that the link is selected; if it is 0, it means that the link is not selected. For an undirected network, there is b ij =b ji ; Define the k×k matrix C as the effective link selection matrix for a certain service path, where the element c ij is the selected link, and it is guaranteed that the link is connected, that is, c ij =a ij ×b ij , if c ij =1, then this link is the routing component of the service, and this link allocates one unit of time slot resources to the corresponding service. For an undirected network, there is c ij =c ji ; Define the normalized cost function F of end-to-end communication service m with the same load (m) for: Among them, m∈[1,N], parameter H (m) Indicates the number of decision path hops for service m. Indicates the number of hops of the shortest path to service m; Among them, the parameter k is the number of network nodes, Select an element in the matrix for the effective link of the path of service m; Step 1.2: Model the routing decision problem of optimal resource allocation for multiple services in the network; Under the condition of limited time slot resources, based on the network selecting the minimum total cost of the path for N services, the performance objective function is formulated as: Set constraints. The constraints of this target optimization problem are path connectivity constraints and node time slot resource constraints.

3. The method for cross-layer joint optimization of wireless ad hoc network routing and resources based on genetic algorithm according to claim 2, characterized in that: In step 1.1, matrix A describes the physical connection status of the network and provides underlying constraints for routing optimization. Matrix B encodes the end-to-end path selection strategy of the service flow, which is independent of the actual network status. Matrix C integrates physical connectivity and service requirements to screen out the actual available routing links. The relationship between them is: C=A⊙B Only when a ij =1 and b ij =1, c ij =1, otherwise 0.

4. The method for cross-layer joint optimization of wireless ad hoc network routing and resources based on genetic algorithm according to claim 2, characterized in that: In step 1.2, the path connectivity constraint is: Among them, s and d are the source node and sink node of a certain service, i,j∈[1,k]; For a source node and a sink node in a path selected by a service, there is only one other node in the path connected to it through a link. For other nodes passed through in the path selected by the service, there must be two nodes in the path connected to it.

5. The method for cross-layer joint optimization of wireless ad hoc network routing and resources based on genetic algorithm according to claim 2, characterized in that: In step 1.2, the node time slot resource constraint is: in It indicates the time slot resource usage of any node j by the paths selected by N services.

6. The method for cross-layer joint optimization of wireless ad hoc network routing and resources based on genetic algorithm according to claim 1, characterized in that: The step 2 is specifically as follows: Step 2.1: Generate a set of candidate paths and calculate the shortest path for each business using the Dijkstra algorithm And generate the first n shortest paths through Yen's algorithm; The algorithm generates a set of candidate paths for each business m. Step 2.2: Define the chromosome encoding rules. The chromosome consists of N genes. Each gene corresponds to a business. The value range is [1, n]. The gene value g m =x indicates that service m selects a path Step 2.3: Design a fitness function to convert F into a positive indicator: Step 2.4: Initialize the population and generate feasible chromosomes through random generation and constraint correction; Step 2.5: Iteratively perform selection, crossover, and mutation operations, and update the population in combination with the elite retention strategy; Step 2.6: Set the maximum number of iterations, output the optimal chromosome, and resolve it into the routing path and time slot allocation plan for each service.

7. The method for cross-layer joint optimization of wireless ad hoc network routing and resources based on genetic algorithm according to claim 6, characterized in that: The step 2.3 is specifically as follows: At the same time, resource constraints need to be handled through penalty items, and the penalty function is defined: If the resource usage of a node exceeds the capacity C, the excess portion max(0,·) is accumulated to ensure it is non-negative; Finally, the weight coefficient β is introduced to balance the importance of the objective and the constraint, and its expression is:

8. The method for cross-layer joint optimization of wireless ad hoc network routing and resources based on genetic algorithm according to claim 6, characterized in that: The step 2.4 is specifically as follows: First, randomly generate and assign a path number g to each chromosome for business m. m ∈[1,n], and then perform constraint modification. If the load of node i Adjust according to the following priorities: Replace the service with the highest load , select a candidate path that does not contain node i. If there is no optional path, randomly disconnect a link associated with node i to force a detour. At the same time, depth-first search is used to detect path breaks and replace broken links with valid links in the candidate paths.

9. The method for cross-layer joint optimization of wireless ad hoc network routing and resources based on genetic algorithm according to claim 6, characterized in that: The step 2.6 is specifically as follows: Map gene values ​​to paths for each business {P (1) ,P (2) ,…,P (N) }, and obtain the time slot allocation scheme for each node based on the generated path set.

10. The method for cross-layer joint optimization of wireless ad hoc network routing and resources based on genetic algorithm according to any one of claims 1 to 9, characterized in that: The method is used for intelligent transportation and vehicle networking, disaster emergency communications, drone networks, the Internet of Things, and mobile ad hoc networks.

Images