Implementation method of AODV routing protocol based on improved HHO and topology change degree

Abstract

The present invention discloses an AODV routing protocol implementation method based on improved HHO and topology change degree. The AODV routing protocol quantifies the topology change degree of nodes by evaluating the change degree of neighbor nodes, the change degree of signal strength and the survival time of neighbor node links. In high-density areas, high-speed nodes only serve as source nodes or target nodes. In low-density areas, all nodes participate in routing forwarding. When the target node receives all path information, the Harris Eagle algorithm HHO is used to select the optimal path and reply information according to the optimal path. The routing protocol of the present invention has achieved remarkable technical effects in key indicators such as throughput, packet loss rate, average end-to-end delay and number of routing requests.

Description

Technical Field

[0001] The invention belongs to the field of network technology, and in particular relates to an AODV routing protocol implementation method based on improved HHO and topology change degree. Background Art

[0002] Routing protocols for mobile ad hoc networks (MANET) can be divided into two categories according to their working mechanisms: table-driven routing protocols and on-demand routing protocols. Table-driven routing protocols maintain routing information to other nodes for each node and keep the validity of the routing table through regular updates. In contrast, on-demand routing protocols only establish routes when needed, thereby reducing unnecessary overhead and energy consumption. As a typical representative of on-demand routing protocols, the Ad hoc On-Demand Distance Vector (AODV) routing protocol is widely used because of its excellent performance in highly dynamic and resource-constrained MANET environments. However, the current AODV routing protocol still has the following shortcomings:

[0003] (1) The AODV routing protocol selects routes based mainly on the number of hops. However, in scenarios with high load, this distance vector-based selection criterion is often not optimal and may lead to uneven distribution of network load. The AODV routing protocol tends to prioritize the shortest path, which in a common square network topology will cause nodes at the center of the topology to face a higher risk of congestion, thereby increasing packet loss and latency. In addition, the selection of the shortest path may pass through some nodes with unstable links, further reducing network performance.

[0004] (2) The AODV routing protocol does not fully consider the scenario of limited node power during the path selection process. If the number of route hops is used as the main selection basis, when some nodes in the network have low power, the protocol may still include these nodes in the shortest path. This situation will cause the node power to be consumed rapidly, leading to node death and affecting the stability of the network.

[0005] (3) In scenarios where topology changes frequently, the AODV routing protocol still uses a flooding mechanism to process routing requests. In a high-density network environment, this mechanism will bring about a large control overhead, resulting in bandwidth waste and exacerbating network congestion problems. Summary of the invention

[0006] In view of the above status of the prior art, the present invention proposes an AODV routing protocol (TD-DBESHHO-WAODV) based on topology change degree and improved Harris Hawk algorithm.

[0007] The present invention adopts the following technical scheme:

[0008] The AODV routing protocol implementation method based on the improved HHO (Harris Hawk algorithm) and topology change degree quantifies the topology change degree of the node by evaluating the neighbor node change degree, signal strength change degree and neighbor node link survival time; in high-density areas, high-speed nodes only serve as source nodes or target nodes; in low-density areas, all nodes participate in routing forwarding; when the target node receives all path information, the Harris Hawk algorithm is used to screen the optimal path and respond according to the optimal path.

[0009] Preferably, the density of the area where the node is located is determined by counting the number of neighbor nodes at a specific moment. If the number of neighbor nodes is greater than a threshold, it indicates that the area where the node is located is a high-density area; otherwise, it is a low-density area.

[0010] Preferably, the node i exist t Calculate the probability of time , the calculation formula is as follows:

[0011] (1);

[0012] In the formula, Representation Node i exist t The number of neighbor nodes at a certain moment, K is a constant.

[0013] As a preference, high-speed nodes calculate the topological change degree of the node itself. TD To judge.

[0014] As a preference, weight , and They are the neighbor node change degrees , signal strength variation and link stability The coefficient of topological change The calculation formula is as follows:

[0015] (2);

[0016] (3).

[0017] As a preference, the neighbor node change degree is specifically as follows:

[0018] nodei exist The formula for the change degree of neighbor nodes within time is:

[0019] (4);

[0020] In the formula, Is a node i exist t The set of neighbor nodes at time, yes i Node in The neighbor node set after time; when the node i exist After time, when all neighbor nodes are updated to new nodes, the neighbor node change degree The value is 1; if After time, the neighbor nodes remain unchanged, and the neighbor node change degree The value is 0;

[0021] The signal strength changes are as follows:

[0022] node i exist The formula for the signal strength change over time is:

[0023] (5);

[0024] (6);

[0025] In the formula, is the neighbor node j of node i in Signal strength value at the moment; Indicates at time t and During this period, it has been a node i The set of neighbor nodes, Is a node i Neighbor nodes j ,exist t Signal strength value over time;

[0026] When the node i exist After a certain time, all the original neighbor nodes lose contact and are replaced by new neighbor nodes. The signal strength changes The value is 1; if After a certain time, if the neighbor nodes do not change and the corresponding signal strength does not change, then the signal strength change degree is The value is 0;

[0027] The link stability is as follows:

[0028] nodei exist T The formula for link stability over time is:

[0029] (7);

[0030] In the formula, T is the link determination period, is T During the cycle, nodes i With neighbor nodes j The link maintenance time between nodes i exist T During the period, the link is broken, and the maintenance time of each link is 0, then the link maintenance time is 0; if T During the period, all links maintain stable connections, and the maintenance time of each link is T , then the link maintenance time is 1.

[0031] Preferably, the quality of the route is evaluated by the congestion degree of the route nodes, the signal quality, the energy consumption speed and the number of route hops.

[0032] Preferably, the Harris Hawk algorithm adds nonlinear escape energy, chaos mapping, dynamic reverse learning and the spiral flight mechanism of the vulture search algorithm.

[0033] The present invention comprehensively evaluates the congestion, signal quality, energy consumption speed and routing hop count of each node to comprehensively measure the quality of the routing, thereby selecting a more robust routing; and dynamically optimizes the routing selection of the WAODV routing protocol by introducing the Harris Hawk Optimization Algorithm HHO to select the most robust routing in the current network. The present invention utilizes a variety of mechanisms including dynamic reverse learning and spiral flight of the vulture search algorithm to overcome the shortcomings of HHO that it is easy to fall into local optimal solutions, thereby improving the performance of the algorithm. The present invention introduces the concept of topological variability to reduce the probability of nodes with a large degree of topological variability becoming intermediate nodes of the routing in high-density areas. After simulation, the results show that the routing protocol of the present invention has achieved significant technical effects on key indicators such as throughput, packet loss rate, average end-to-end delay and number of routing requests. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is the route discovery flow chart of AODV routing protocol;

[0035] Figure 2 It is a comparison chart of escape energy curve;

[0036] Figure 3 It is a time series comparison chart of escape energy;

[0037] Figure 4 It is a logistic mapping situation diagram;

[0038] Figure 5 It is a schematic diagram of the reverse solution position;

[0039] Figure 6 It is a schematic diagram of the location of QOBL and QRBL;

[0040] Figure 7 It is the flow chart of DBESHHO algorithm;

[0041] Figure 8 It is the flow chart of AODV routing protocol based on improved HHO;

[0042] Fig. 9 It is a flow chart of route discovery based on the improved HHO and topology change degree AODV (TD-DBESHHO-WAODV) routing protocol;

[0043] Fig.10 This is a graph showing the impact of node speed on throughput and packet loss rate;

[0044] Fig.11 It is a graph showing the effect of node speed on average end-to-end delay and number of routing requests;

[0045] Fig.12 This is a graph showing the impact of the number of nodes on throughput and packet loss rate;

[0046] Fig.13 It is the effect of the number of nodes on the average end-to-end delay and the number of routing requests;

[0047] Fig.14 It is the Harris Eagle algorithm flow chart. DETAILED DESCRIPTION

[0048] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings and technical principles.

[0049] 1. Weighted AODV routing protocol optimization method

[0050] To solve the problem of unstable routing selection, this embodiment proposes a routing protocol based on weighted optimization AODV (WAODV). Different from the traditional method that only relies on the order of information arrival and the number of routing hops, this embodiment evaluates the routing by incorporating more key factors, and evaluates the overall quality of the routing through the congestion of routing nodes, signal quality, energy consumption speed and number of routing hops, so as to select a more robust routing with higher quality.

[0051] This embodiment uses a threshold: RSSI_TH, which represents the threshold of signal quality. When the signal strength of the route request information received by the node is lower than this threshold, the request message is ignored to avoid affecting the data transmission quality after route generation.

[0052] When a node needs to send data to the destination node, it will select a path to send according to the information in the local routing table. If there is no path to the destination node in the routing table, the source node will initiate a route request process and broadcast an RREQ (Route Request) message. If there is no route to the destination node, the source node initiates a route request process and broadcasts an RREQ message. The intermediate node that receives the RREQ information calculates the signal quality of the link based on the signal transmission between the two nodes. If the signal quality is less than the threshold, the request information is discarded. If the signal strength meets the requirements, a reverse route to the source node is established, and the hop count and signal quality are updated to the reverse route entry. At the same time, the congestion degree and energy consumption speed information of this intermediate node are attached to the RREQ message, and this information is calculated in advance according to the operating conditions of the node and updated in real time. After modifying the RREQ message, this intermediate node re-broadcasts it. All intermediate nodes follow this process until the route request reaches the destination node. When the destination node receives the route request, it waits for a specific time to receive all the route request information. The destination node evaluates all the routes by combining the hop count, signal quality of all the routes, and the congestion degree and energy consumption speed of the intermediate nodes of the route, and selects the route with the best weighted score of the route, and sends the RREP (Route Reply) message to the source node through this route. After receiving the reply message, the intermediate node will update or establish a forward route to the destination node according to this message and clear the redundant items in the routing table. This process continues until the route reply message returns to the source node. After receiving the reply message, the source node successfully establishes a forward route to the destination node and then sends data through this path.

[0053] 1.1. Weighted Information

[0054] When considering selecting a certain route, the first-arrived route request information is not necessarily the best. Some intermediate nodes are limited by poor signal and high congestion degree, and the data transmission effect is not good. Therefore, this embodiment comprehensively considers the congestion degree, signal quality, energy consumption speed and route hop count of each node to calculate the weighted score of the route.

[0055] (1) Congestion Degree

[0056] The congestion of a node can be determined by the load ratio of the queue buffer. To measure the degree of congestion, the actual load of the queue (average queue length) can be compared with the processing capacity of the queue (queue capacity) to obtain a percentage. This percentage represents the ratio of the actual load of the queue to its maximum carrying capacity. In time period t, the congestion of the node It can be expressed by the following formula:

[0057] (1);

[0058] Where C is the capacity of the queue, It's time point τ The queue length at the time.

[0059] (2) Signal quality

[0060] The signal quality of the route determines the stability of data transmission, and it is less likely to cause a link failure. RSSI The following formula can be used for calculation:

[0061] (2);

[0062] In the formula, is the receiving antenna gain, is the transmit antenna gain, is the transmission power, f is the frequency used for wireless network communication, and c is the speed of light.

[0063] According to the signal quality between two nodes, the threshold RSSI_TH of the signal quality is calculated. The formula is as follows:

[0064] (3);

[0065] In the formula, R Indicates the range of the antenna. If the node is RSSI Less than RSSI_TH If the threshold is exceeded, all messages from the neighboring node are discarded.

[0066] (3) Energy consumption rate

[0067] By calculating the energy consumption of the node in different working modes, a comprehensive energy consumption rate indicator is provided. Since wireless nodes will switch between multiple modes in actual operation, it is very important to consider the power consumption and time allocation in different modes. The specific calculation formula is as follows:

[0068] (4);

[0069] In the formula, is the energy consumed per bit of data sent, is the length of the data sent, is the energy consumed per bit of data received, is the length of the received data, is the power consumption in idle monitoring mode, is the time in idle monitoring mode, is the detection cycle.

[0070] (4) Route hop count

[0071] The number of hops required for each message to be delivered from the source node to the destination node.

[0072] (5) Weighted Route Scoring

[0073] For each route, its weighted score RWS is calculated based on its total number of hops, total signal quality, congestion of intermediate nodes, energy consumption rate, etc. The quality of the route is judged by the score. The smaller the value, the better the route. The specific calculation formula is as follows:

[0074] (5);

[0075] In the formula, , , and They are the weights of routing hops, signal quality, congestion, and energy consumption speed. ; is the total number of hops in the route, It is i The congestion degree of the intermediate nodes, It is i The energy consumption rate of each node, The signal quality of each route RSSI The sum.

[0076] 1.2. WAODV routing protocol message format

[0077] This embodiment modifies the format of RREQ, RREP messages and local routing tables in the AODV routing protocol. By modifying the RREQ message, additional information is added to the routing information so that the cost of each path can be analyzed at the target node to determine its quality. By modifying the RREP and local routing table information, it is convenient for the RREP message to correctly find the path back to the source node.

[0078] (1) RREQ message modification

[0079] The Total Signal Quality field is added to record the total signal quality of the route; the Congestion Degree field is used to record the node congestion degree; and the Energy Consumption Rate field is used to record the node energy consumption rate. The total number of records is calculated by HopCount. The modified RREQ message format is shown in Table 1:

[0080] Table 1 Part of the RREQ message format of the WAODV routing protocol

[0081]

[0082] In Table 1, the ellipsis part indicates the congestion degree and energy consumption rate of each intermediate node attached later (the length of the message is determined by the number of node hops).

[0083] (2) RREP message modification

[0084] The RREP message format is modified by adding the Signal Quality and Total Signal Quality fields. The Total Signal Quality field is used to find the forward route. The modified RREP message format is shown in Table 2:

[0085] Table 2 Partial RREP message format of WAODV routing protocol

[0086]

[0087] (3) Description of special fields in the HELLO (active route detection) message

[0088] The message format of HELLO is the same as the modified RREP message format. In the HELLO message, the node uses the SignalQuality field to indicate the signal quality value of the other party's message; the Total Signal Quality field indicates the signal quality value determined between the two. This value is not modified during the route discovery process. Other fields are the same as the AODV routing protocol.

[0089] (4) Routing table modification

[0090] A new Total Signal Quality field is added to the local routing table to record the total signal quality from the current node to the destination node. The modified routing table entry is shown in Table 3:

[0091] Table 3 Partial routing table message format of WAODV routing protocol

[0092]

[0093] 1.3. WAODV routing protocol routing discovery process

[0094] The WAODV routing protocol mainly includes five links, namely routing initialization, RREQ message forwarding, routing request reception completion, RREP message forwarding, and routing response reception completion. The details are as follows:

[0095] (1) Routing initialization

[0096] When a node starts running, it first calculates the congestion and energy consumption rate of its own node and stores the calculated values ​​locally. The signal quality between the node and the neighboring node is determined through the HELLO message. If the signal quality is less than the RSSI_TH threshold, the two nodes do not receive information from each other. If the signal quality is greater than the threshold, a routing table entry for the neighboring node is created and the determined signal quality is stored in the corresponding field.

[0097] (2) RREQ message forwarding

[0098] The source node initiates a route request to find a route and broadcasts a RREQ message, in which the value of the Total SignalQuality field is initially 0.

[0099] When a neighbor node receives a message, it first checks whether it is the destination node of the RREQ message. If not, it determines whether the signal quality between itself and the previous node is greater than RSSI_TH If the signal quality of the neighbor node is less than RSSI_TH , then discard the RREQ message, otherwise establish a reverse route to the source node. The establishment of the reverse route requires the signal quality value between the node and the previous node, plus the signal quality value in the RREQ message, to be stored in the TotalSignalQuality field in the reverse routing table, representing the total signal quality value from the source node to the node, and replace the corresponding value in the RREQ message with this value. After the reverse route is established, increase the hop count field of the RREQ message by one, attach the congestion degree and energy consumption rate of the node, and then the node rebroadcasts the route request.

[0100] (3) Routing request reception completed

[0101] When the destination node receives the RREQ message, it also first establishes a reverse route to the source node, calculates the total signal quality value of the route, and stores it in the corresponding routing table entry. After waiting for a period of time, the destination node receives multiple RREQ messages, calculates the weighted costs of different routes, obtains the route with the lowest cost, and uses this route to update the routing table to the source node. The RREP message is unicasted to the source node that initiated the routing request through this route. The Signal Quality field value in the message is 0, and the Total Signal Quality field value is the total signal quality value of the optimal route.

[0102] (4) RREP message forwarding

[0103] When the intermediate node receives the RREP message, it calculates the total signal quality value of the propagated route, establishes the corresponding forward route based on the value, and updates the Signal Quality field in the RREP message. Based on the value of TotalSignal Quality minus Signal Quality in the RREP message, it searches for the correct reverse route locally and deletes redundant routes. It then continues to unicast the RREP message to the next node according to the routing table.

[0104] (5) Routing response reception completed

[0105] When the source node (destination node of the RREP message) receives the RREP message, it calculates the total signal quality of the route, creates a forward route to the destination node (source node of the RREP message), and updates the Total Signal Quality field of the routing table. At this point, the source node and the destination node can communicate with each other according to the routing table.

[0106] The complete routing process is as follows Figure 1 shown.

[0107] 2. AODV routing protocol optimization method based on improved Harris Hawk algorithm

[0108] The AODV routing protocol based on weighted optimization improves network performance by comprehensively evaluating four factors: congestion, signal quality, energy consumption speed, and number of routing hops. The optimization effect depends on the pre-set weights to select relatively high-quality routes. To further improve performance, it is allowed to increase energy consumption appropriately and sacrifice energy balance in exchange for better network performance.

[0109] To solve the above problems, this embodiment introduces the Harris Hawks Optimization (HHO) algorithm, which uses it to dynamically adjust weights to cope with dynamic changes in the network in a more robust way. HHO simulates the hunting behavior of an eagle. After the AODV routing protocol receives all routing requests, it flexibly adjusts the weights according to performance indicators such as throughput, packet loss rate, and average end-to-end delay, and optimizes the routing cost calculation method in real time. This method ensures that routing selection can adapt to different network conditions while allowing appropriate increase in energy consumption and moderate sacrifice of energy balance, thereby achieving more efficient routing selection in complex environments, effectively improving network throughput, reducing packet loss rate and average end-to-end delay, and improving network performance.

[0110] This embodiment proposes a fusion HHO algorithm based on dynamic reverse learning (DBESHHO), which adds nonlinear escape energy, chaos mapping, dynamic reverse learning, and the spiral flight mechanism of the vulture search algorithm. Through these improvements, the global exploration and local optimization are better balanced, the risk of premature convergence is reduced, and the performance of the Harris Hawk algorithm in complex optimization problems is improved.

[0111] 2.1 Nonlinear escape energy coefficient

[0112] In the traditional Harris Eagle algorithm, the key factor that determines whether the algorithm performs a global search or a local search is the escape energy coefficient E. This coefficient has an important impact on the optimization ability of the algorithm. Usually, in the early stages of the algorithm's iteration, it is hoped that the algorithm will perform more global searches to ensure that potential optimal solutions are not missed. This is because global searches help explore the entire search space and avoid falling into local optimality. However, as the iterations progress, in order to speed up convergence and find the optimal solution as quickly as possible, it is hoped that the algorithm will perform more local searches to accurately adjust and optimize the solution. Therefore, the DBESHHO algorithm uses a nonlinear factor as the new escape energy coefficient, and the formula is as follows:

[0113] (6);

[0114] In the formula, represents the initial escape energy of the rabbit, which is a random number uniformly distributed in the interval (-1,1). is the current iteration number, is the maximum number of iterations. The comparison between the traditional escape energy coefficient HHO and the improved escape energy coefficient DBESHHO is shown in Figure 2 and Figure 3 shown.

[0115] This formula ensures that the escape energy coefficient changes nonlinearly during the algorithm operation, so that the algorithm has a higher global search capability in the early stage, and gradually turns to a stronger local search capability in the later stage, thereby improving the optimization efficiency and solution accuracy. Through this nonlinear adjustment strategy, the DBESHHO algorithm can effectively balance global exploration and local optimization, and enhance the performance of the algorithm in complex problems. Specifically, the dynamic adjustment of the escape energy coefficient enables the algorithm to more flexibly adapt to the optimization needs of different stages during the search process, thereby improving the overall optimization ability of the algorithm.

[0116] 2.2 Chaotic Mapping

[0117] Chaotic mapping is a common phenomenon in nonlinear dynamic systems, showing complex behavioral characteristics between randomness and regularity, with characteristics such as boundary, randomness, regularity and ergodicity. These characteristics make chaotic mapping play an important role in understanding and studying complex systems. At present, a variety of chaotic mapping (Logistic mapping) has been applied to intelligent algorithms, among which Logistic mapping is a common method. Although the Logistic mapping function form is simple, its dynamic behavior shows complex and unpredictable characteristics within a specific parameter range.

[0118] (7).

[0119] In the formula, The value of is a constant, usually in the range [1,4]. It is in n The value of the iteration.

[0120] Logistic mapping can generate random sequences, and the results are affected by The impact of In the range [3.6,4], x The value of shows chaotic characteristics in the interval (0,1), which is exactly the expected result. The Logistic mapping value is as follows Figure 4 shown.

[0121] When improving the HHO algorithm, chaotic mapping is used to initialize the population position. Compared with the use of random operators, the chaotic mapping operator can generate a more evenly distributed and non-repetitive initial population, thereby improving the efficiency and accuracy of the algorithm. Specifically, when the population position is initialized using the Logistic mapping method, each dimension of each solution vector generates a corresponding initial value through chaotic mapping. In this way, the entire initial population can cover the search space more evenly, thus laying a good foundation for the subsequent optimization process.

[0122] 2.3 Dynamic Reverse Learning

[0123] In metaheuristic optimization algorithms, population initialization is usually achieved by randomly generating individuals in the search space. Although this random initialization can ensure that the distribution of the population in the search space is relatively dispersed, it is difficult to ensure the quality of the initial solution. Studies have shown that the quality of initialization directly affects the convergence speed of the algorithm and the accuracy of the final solution. Therefore, in order to improve the performance of the algorithm, this embodiment introduces a dynamic reverse learning strategy (Dynamic Opposition-based Learning, DOBL).

[0124] Dynamic opposition learning is an emerging opposition learning strategy (Opposition-based Learning, OBL), which is inspired by random opposition-based learning (ROBL), quasi-opposite based learning (QOBL) and quasi-reflection based learning (QRBL).

[0125] For a solution X in the search space, its reverse solution is defined as follows:

[0126] (8);

[0127] In the formula, lb is the lower bound of the search space, ub is the upper limit of the search space. The specific schematic diagram is as follows Figure 5 shown.

[0128] Based on OBL, ROBL introduces a factor randomly distributed between (0,1) , reverse solution It is no longer a fixed symmetric point, but a solution randomly distributed within a range. This can increase the diversity of solutions and prevent the algorithm from falling into a local optimum. Its definition is as follows:

[0129] (9);

[0130] QOBL and QRBL extend the solution to [ , ], the midpoint of the search space is added for reference, making the generated solution more accurate, and its quasi-reverse number and quasi-reflection number is defined as follows:

[0131] (10);

[0132] (11);

[0133] In the formula, and are random numbers uniformly distributed within their respective ranges, with values ​​ranging from Figure 6 shown.

[0134] Compared with the traditional reverse learning method, the dynamic reverse learning strategy enhances the algorithm's search capability by expanding the search range of the current solution and its symmetric solution to the entire search space. This extension allows the search space to be dynamically adjusted, ensuring that the population always keeps exploring for better solutions during the evolution process. The mathematical definition is as follows:

[0135] (12);

[0136] In the formula, w is the dynamically adjusted weight factor, ( i =1,2) is a random number uniformly distributed between (0,1).

[0137] if Out of bounds, you can lb , ub ] to generate a random number to get a new This approach has two main advantages: on the one hand, the randomly generated reverse solutions expand the search space and increase the possibility of the algorithm converging to the global optimal solution, thereby enhancing the development ability; on the other hand, since the search space is dynamically changing, the DOBL strategy performs well in exploration ability. At the same time, the weight factor w The adjustment of enables the generation of reverse solutions to dynamically adapt to changes in the search space. At different stages of the search, w The dynamic adjustment of helps the algorithm to conduct extensive searches in the early stage and then gradually converge to a better area, achieving a smooth transition from global exploration to local development.

[0138] By applying the DOBL strategy, not only the diversity of candidate solutions is further enriched on the basis of ROBL, QOBL and QRBL strategies, and the quality of solutions is improved, but also it can effectively avoid falling into local optimal solutions and accelerate the convergence speed, making it more adaptable and robust in complex optimization problems.

[0139] Therefore, the DOBL strategy is introduced to improve the HHO algorithm. The specific execution steps are: Step 1, calculate the dynamic reverse learning population according to formula (12); Step 2, merge the dynamic reverse learning population with the original population to form a larger population; Step 3, sort the fitness values ​​of the current eagle population; Step 4, select the top N individuals with better fitness as the new generation population to improve the population quality of the HHO algorithm. In this paper, the DOBL strategy will be applied in two stages of HHO: the first stage is to use the DOBL strategy during the initialization process of the HHO algorithm to increase the diversity of the population, thereby actively affecting the global search; the second stage is to use the DOBL strategy in each iteration (as shown below, step 11, when the cutoff condition of the iteration is not met, continue to execute step 3 and use the dynamic reverse learning strategy) to generate a reverse solution and guide the individual to jump out of the local optimal solution.

[0140] 2.4. Fusion Vulture Search Algorithm

[0141] The Bald Eagle Search (BES) algorithm is a swarm intelligence optimization algorithm proposed by Alsattaretal in 2020. The BES algorithm is a swarm intelligence algorithm created by simulating the habits of vultures after studying their hunting habits. The Bald Eagle Search Algorithm has excellent convergence ability and is divided into three stages: selecting search space, searching for prey in space, and diving to capture prey.

[0142] (1) Selecting the search space

[0143] The vulture performs random searches within the range, and selects the search area based on the location of the prey and its own experience. The vulture's position update formula is:

[0144] (13);

[0145] In the formula, It is the position with the highest fitness among the current vulture individuals. is the parameter that controls the position change, and its value range is (1.5,2). is a random number uniformly distributed in the range (0,1). is the average position of all vultures, It is i Only the location of the vulture.

[0146] (2) Searching for prey in space

[0147] The vulture selects a spatial location to search for prey, flies in a spiral shape to search for prey, and accelerates the search speed to find the best dive capture position. The formula for updating the flight position using polar coordinates is:

[0148] (14);

[0149] (15);

[0150] (16);

[0151] (17);

[0152] In the formula, Indicates that before the population is updated, i +1 vulture position, and represents the coordinates of the vulture in the plane rectangular coordinate system, is the polar radius of the polar coordinates, represents the polar angle of polar coordinates, and are the parameters that control the vulture's spiral search trajectory, and their ranges are (0,5), (0.5,2), ( i =2,3) is a random number uniformly distributed in the range (0,1). x controls the horizontal position update of the vulture in the polar coordinate system; y controls the vertical position update of the vulture in the polar coordinate system; r is a random variable related to the search direction and amplitude, which is determined by the polar angle θ(i) and a constant factor R and another random variable to update.

[0153] (3) Diving to catch prey

[0154] The vulture will quickly dive towards the target prey from the optimal position in the search space. At the same time, other individuals in the population will also move towards this optimal position and jointly attack. Their motion state is still described by the polar coordinate equation, and the position update formula is:

[0155] (18);

[0156] (19);

[0157] (20);

[0158] (twenty one);

[0159] In the formula, , Indicates the intensity of the vulture's dive toward the optimal point and the center point, with a value range of (1,2). is a random number uniformly distributed in the range (0,1). i =4,5; sinh represents the hyperbolic sine function; cosh represents the hyperbolic cosine function;

[0160] The vulture search algorithm has good convergence ability, and its ability to jump out of the local optimal solution mainly depends on the spiral flight in the prey stage of the search space. The introduction of the spiral flight strategy brings many improvements to the algorithm. First, spiral flight expands the search space, allowing individuals to conduct more extensive exploration, thereby enhancing the global search ability. Secondly, the complex trajectory of this strategy increases the diversity of the search, effectively avoiding the algorithm from falling into the dilemma of local optimality. In addition, the spiral flight strategy can be flexibly adjusted at different stages of the algorithm, so that the algorithm can strike a balance between global exploration and local development. Spiral flight not only imitates the hunting behavior of predators in nature, but also improves the adaptability of the algorithm to a variety of complex optimization problems, thereby improving the overall optimization performance.

[0161] In the Harris Hawk algorithm, the soft encirclement strategy formula of the Harris Hawk is changed, the spiral flight strategy of the vulture search algorithm is introduced, and formula (14) is used instead, and the modified formula (17) is as follows:

[0162] (twenty two);

[0163] In the formula, The value is 1.5. is the current iteration number, is the maximum number of iterations, ( i =5,6) is a random number uniformly distributed in the range (0,1).

[0164] In the original spiral flight strategy, the parameters R The value range of is fixed to a value between (0,5). However, after modification, R It is no longer a fixed value, but is used Instead, it is a value that decreases with iteration. In this way, the flight area will gradually shrink, thereby accelerating the convergence of the algorithm. This improvement aims to reduce the search space and focus on exploring in areas where the optimal solution is more likely to be found, thereby improving the overall efficiency and performance of the algorithm.

[0165] 2.5. The DBESHHO algorithm combines multiple improvement strategies to achieve efficient global search and local development. The specific operation steps of the algorithm are as follows:

[0166] Step 1: Enter the number of eagles N , maximum number of iterations , the upper and lower boundaries of each dimension of the search space ub , lb ;

[0167] Step 2: Initialize the eagle group according to formula (7) ;

[0168] Step 3: Update the position of the eagle group according to the dynamic learning mechanism of formula (12);

[0169] Step 4: Calculate the fitness function and convert the best fitness value , denoted as ;

[0170] Step 5: Update the nonlinear escape energy according to formula (6);

[0171] Step 6: If the energy escapes , then update the eagle group position according to formula (38):

[0172] Step 7: If And the probability of escape , then update the eagle group position according to formula (14) and (22);

[0173] Step 8: If and , then update the eagle group position according to formula (43);

[0174] Step 9: If and , then update the eagle group position according to formula (47);

[0175] Step 10: If and , then update the position of the eagle group according to formula (42);

[0176] Step 11: If If satisfied, continue to step 3. Otherwise, the program ends and outputs the optimal solution and optimal function value.

[0177] The algorithm flow of DBESHHO is as follows Figure 7 shown.

[0178] 2.6. Fitness Function of Routing Protocol

[0179] The fitness function is an important indicator for evaluating the quality of solutions in optimization algorithms. It is used to measure the closeness of the solution to the problem goal. The fitness function is usually designed according to the specific requirements of the problem, and the goal is to optimize the fitness (maximize or minimize) through optimization.

[0180] In the AODV routing protocol optimization, the fitness function helps select a routing path with better performance (high throughput, low packet loss rate and low latency) by quantifying the expected throughput, packet loss rate and latency. Adding these expected indicators to the fitness function makes the optimization process more clearly targeted at the key goals of network performance, thereby achieving an effective trade-off between multiple performance indicators. The fitness function can better reflect the status of network transmission and avoid the deviations and limitations that may be caused by minimizing a single goal.

[0181] The fitness function F Defined as:

[0182] (twenty three);

[0183] in, is the weighted cost of the route, calculated as shown in formula (5), represents the normalized expected throughput, represents the normalized expected packet loss rate, Represents the normalized expected delay.

[0184] (1) Expected throughput

[0185] Throughput is affected by network congestion and signal quality. The theoretical maximum throughput is used as a benchmark and adjusted according to the total signal strength of the route, the number of hops, and the congestion of each node to obtain an expected throughput that is close to reality. The evaluation calculation formula is as follows:

[0186] (twenty four);

[0187] In the formula, represents the theoretical maximum throughput, is the total number of hops in the route, is the total signal quality of the route, It is i The congestion level of a node.

[0188] (2) Expected packet loss rate

[0189] The packet loss rate can be estimated by congestion, total signal quality, and number of hops. Links with poor signal quality usually result in higher packet loss rates, while severely congested nodes increase the possibility of packet loss, and an increase in the number of hops brings potential packet loss risks. Taking these factors into consideration, the estimation of the packet loss rate can more accurately reflect the transmission stability in the network environment. The evaluation calculation formula is as follows:

[0190] (25);

[0191] (3) Expected delay

[0192] The delay of a single route can be expressed as the sum of the delays of each hop. Assuming that the delay of each node mainly depends on the link transmission delay and processing delay, the total delay of the route can be estimated based on the number of hops and congestion. The evaluation calculation formula is as follows:

[0193] (26);

[0194] In the formula, It is i The bandwidth of each node.

[0195] (4) Normalization calculation

[0196] The normalized calculation methods of expected throughput, expected packet loss rate and expected delay are:

[0197] (27);

[0198] (28);

[0199] (29);

[0200] Wherein, the max subscript represents the maximum value of the corresponding expected indicator among all received routes, and the min subscript represents the minimum value of the corresponding expected indicator among all received routes.

[0201] The overall process of AODV routing protocol based on improved HHO is as follows Figure 8 shown.

[0202] 3. AODV routing protocol optimization based on improved HHO and topology change

[0203] 3.1 Overview of optimization methods based on topological change

[0204] In a high-density node environment, there are multiple similar routes in the network. The nodes on these routes often move at different speeds, especially nodes that move at high speeds are prone to link breakage. Therefore, in high-density areas, selecting nodes with lower moving speeds as intermediate nodes can improve the robustness of routing, reduce the probability of route interruption, and reduce the need to re-initiate route discovery. This strategy helps reduce unnecessary communication overhead.

[0205] In view of the above problems, this embodiment proposes an AODV routing protocol based on improved HHO and topology variation degree (TD), referred to as TD-DBESHHO-WAODV. The protocol quantifies the topology variation degree of the node by comprehensively evaluating the variation degree of neighbor nodes, the variation degree of signal strength, and the survival time of neighbor node links. Specifically, the topology variation degree is used to evaluate the movement speed and topological stability of the node. In high-density areas, if a node is evaluated as a high-speed node, the node will no longer forward RREQ messages, nor will it serve as an intermediate node for other routes. Such high-speed nodes only serve as source nodes or target nodes to avoid frequent link breaks due to high-speed movement.

[0206] In low-density areas, if high-speed nodes do not participate in the role of intermediate nodes, message delivery may be blocked. Therefore, in high-density areas, the probability of judging nodes as high-speed nodes is increased, and the chance of them being intermediate nodes is reduced, while in low-density areas, all nodes are allowed to participate in routing forwarding to ensure network connectivity.

[0207] The density of the area where the node is located is determined by counting the number of neighbor nodes at a certain moment. A large number of neighbor nodes usually means that the node is in a high-density area. A new parameter is introduced, namely ,node i exist t Calculate the probability of time , the calculation formula is as follows:

[0208] (30);

[0209] In the formula, Representation Node i exist t The number of neighbor nodes at a certain moment, K is a constant.

[0210] The high-speed node is determined by calculating the topological change degree of the node itself. TD To proceed (first calculate the probability of topological change, if the random number generated by the current node is less than the calculated probability, then calculate the topological change, if the topological change is greater than the threshold, it is determined to be a high-speed node). TD_TH , the RREQ message will no longer be forwarded. The route discovery process of the TD-DBESHHO-WAODV routing protocol is as follows: Fig. 9 shown.

[0211] If in t Moment, Node i The number of neighbor nodes is less than or equal to K( ), then the node iIn the low-density area, the forwarding times and probability of the RREQ message are not reduced; otherwise, the node is considered i to be in the high-density area ( ), and the probability of forwarding the RREQ message is reduced. Therefore, when a node in the high-density area receives an RREQ message, it first calculates the probability . Subsequently, the node generates a random number p uniformly distributed in the interval (0, 1), and compares it with . If , that is, the forwarding condition is satisfied, the node calculates its own topological change degree TD . If , the node discards the received RREQ message; otherwise, it forwards the received RREQ message.

[0212] The core idea of the TD-DBESHHO-WAODV routing protocol is to preferentially select low-speed and stable nodes as intermediate nodes through the topological change degree, thereby constructing a more robust routing, reducing the risk of link breakage, and improving the performance of the network in high-density and highly changing topological scenarios.

[0213] 3.2. Definition and calculation of topological change degree

[0214] The topological change degree selects the neighbor node change degree, signal strength change degree, and link stability degree as the factors to consider the node movement speed, comprehensively evaluates the node movement situation, and provides a basis for the final robust routing decision.

[0215] (1) Neighbor node change degree

[0216] The neighbor node change degree is used to measure the change of the neighbor node set of a node within a certain period of time, reflecting the stability of the node in the local network. The node compares the neighbor node sets of adjacent cycles to obtain the neighbor node change degree of the node. The node i at time, the formula for the neighbor node change degree is:

[0217] (31);

[0218] In the formula, is the neighbor node set of node i at t moment, is i the neighbor node set of the node after time. When all neighbor nodes of node i at time are updated to new nodes, the neighbor node change degree value is 1. If at After time, the neighbor nodes remain unchanged, and the neighbor node change degree The value is 0.

[0219] (2) Signal strength variation

[0220] The signal strength variation measures the node's moving speed by analyzing the changes in the received signal strength between nodes in two adjacent time periods. The fluctuation of signal strength reflects the change in the distance between nodes. The signal strength variation can be used to determine the movement of nodes in physical locations and further evaluate the reliability of the link. i exist The formula for the signal strength change over time is:

[0221] (32);

[0222] (33);

[0223] In the formula, It's in time t and During this period, it has been a node i The set of neighbor nodes, Is a node i Neighbor nodes j ,exist t The signal strength value at the time. i exist After a certain time, all the original neighbor nodes lose contact and are replaced by new neighbor nodes. The signal strength changes The value is 1. After a certain time, if the neighbor nodes do not change and the corresponding signal strength does not change, then the signal strength change degree is The value is 0.

[0224] (3) Link stability

[0225] Link stability measures the continuity of links between nodes and is one of the important indicators for evaluating node movement speed. When link stability is low, it means that the link is disconnected more frequently, which usually reflects that the node moves faster and the link reliability is poor. On the contrary, high link stability indicates that the link remains connected for a longer period of time, the node moves slower, and the network topology is relatively stable. Therefore, link stability can effectively reflect the movement speed of the node and the risk of link interruption. i exist T The formula for link stability over time is:

[0226] (34);

[0227] In the formula, Tis the link determination period, is T During the cycle, nodes i With neighbor nodes j The link maintenance time between nodes. i exist T During this period, the link is immediately broken, and the maintenance time of each link is close to 0, so the link maintenance time is close to 0. But if T During the period, all links maintain stable connections, and the maintenance time of each link is T , then the link maintenance time Close to 1.

[0228] In order to comprehensively evaluate the mobility of nodes, the topology change TD The calculation is based on three indicators: neighbor node change, signal strength change, and link stability. The weighted method is used to determine the topology change, and these three indicators are linearly combined with appropriate weights. Set the weight , and They correspond to the influence of neighbor node change, signal strength change and link stability respectively. TD The calculation formula is as follows:

[0229] (35);

[0230] (36);

[0231] The uniform weight reflects the impact of these three factors on the topology, indicating that their importance in evaluating the node topology change is relatively consistent, thus ensuring that the comprehensive evaluation is not biased towards a specific factor. The topology change calculated based on these three indicators can more effectively evaluate the node's movement speed and its impact on the network topology, and provide a reference for routing decisions.

[0232] 4. Simulation and results analysis of TD-DBESHHO-WAODV routing protocol

[0233] In order to verify the effectiveness of the TD-DBESHHO-WAODV routing protocol, a comparative test was carried out based on the MATLAB platform, and the performance of the TD-DBESHHO-WAODV routing protocol was compared with the DBESHHO-WAODV, UVACO-AODV, and AODV routing protocols in terms of throughput, packet loss rate, average end-to-end delay, and number of routing requests. Each routing protocol was simulated 10 times independently, and the results were averaged.

[0234] 4.1 Simulation scene settings

[0235] A total of 50 nodes are set within a rectangular range of 1000m×1000m, and the radio communication range of each node is 300m. The movement of the nodes follows the RW mobility model. The maximum movement speed of the nodes in this scenario is increased to 50m / s to simulate a highly dynamic network environment. The nodes randomly select target locations in the area and move toward the target locations at random speeds, with the speeds evenly distributed between 0 and 50m / s. After reaching the target location, the node no longer stays, but selects a new target location to continue moving, causing the node to exhibit an irregular continuous motion pattern. During the communication process, each node will randomly select a communication partner and transmit data once per second, with each transmission containing 4000bits of data, simulating data transmission behavior in a highly dynamic environment. The setting of this high-speed mobile scenario is intended to evaluate the adaptability and performance of different routing protocols in a complex and rapidly changing network environment. The specific simulation parameters are shown in Table 4 below:

[0236] Table 4 Simulation environment parameters

[0237]

[0238] 4.2 Simulation results and analysis

[0239] The performance of the proposed TD-DBESHHO-WAODV routing protocol is compared with that of DBESHHO-WAODV, UVACO-AODV, and AODV routing protocols. The performance of these protocols is evaluated in scenarios with different node speeds and different numbers of nodes.

[0240] (1) Different node speeds

[0241] Fig.10 , Fig.11 The effect of different node speeds on throughput, packet loss rate, average end-to-end delay, and number of routing requests is shown.

[0242] like Fig.10 As shown in the upper middle part of the figure, when the node speed is low, the throughput of the TD-DBESHHO-WAODV routing protocol and the DBESHHO-WAODV routing protocol is not much different, because the topological change degree of the node has not reached the threshold, the node will not discard the RREQ message, and the network performance will not be affected. When the node speed gradually increases, the forwarding of the RREQ message is restricted, which effectively reduces the network congestion and improves the throughput.

[0243] like Fig.10As shown in the figure in the lower middle part, the advantages of the TD-DBESHHO-WAODV routing protocol gradually become apparent as the node speed increases. By reducing the probability of high-speed nodes acting as intermediate nodes in the route, it effectively avoids link breakage and reduces packet loss caused by route interruption. This mechanism ensures that even in high-speed mobile scenarios, the network still maintains a relatively stable route, improving the reliability of data transmission.

[0244] like Fig.11 As shown in the upper middle part of the figure, in the high-speed mobile scenario, the average end-to-end delay of the TD-DBESHHO-WAODV routing protocol is slightly lower than that of the DBESHHO-WAODV routing protocol. This is because the TD-DBESHHO-WAODV routing protocol avoids frequent routing reconstruction processes by reducing the probability of high-speed nodes becoming intermediate nodes in the route. Since high-speed nodes are often prone to link interruptions, resulting in additional delay overhead, TD-DBESHHO-WAODV maintains routing continuity by giving priority to stable nodes, thereby reducing the average end-to-end delay of the network. In scenarios with lower mobile speeds, the average end-to-end delay of the TD-DBESHHO-WAODV routing protocol is slightly higher than that of the DBESHHO-WAODV routing protocol. This is because the additional optimization strategy increases the computational overhead of route selection, resulting in a certain increase in delay.

[0245] like Fig.11 As shown in the lower middle figure, the number of routing requests for all routing protocols increases with the increase of node movement speed. This is because high-speed moving nodes easily make the link unstable, causing frequent routing breaks, resulting in more routing requests. The number of routing requests for the TD-DBESHHO-WAODV routing protocol is less than that of the AODV routing protocol because its routing selection strategy effectively reduces the link breaks caused by high-speed nodes and reduces the need to re-initiate routing requests.

[0246] (2) Different number of nodes

[0247] Fig.12 , Fig.13 The effect of different numbers of nodes on throughput, packet loss rate, average end-to-end delay, and number of routing requests is shown.

[0248] like Fig.12 As shown in the upper middle figure, the throughput of the TD-DBESHHO-WAODV routing protocol is slightly higher than that of the DBESHHO-WAODV routing protocol, but the throughput of both is better than that of the AODV protocol. This is because as the node density increases, the number of nodes involved in judging the degree of topology change increases, which reduces the forwarding frequency of RREQ messages and reduces network congestion.

[0249] like Fig.12 As shown in the lower middle part of the figure, when the node density is large, TD-DBESHHO-WAODV reduces the routing burden of nodes and reduces the packet loss rate by reducing the routing participation probability of high-speed nodes in high-density areas, making it perform better than the DBESHHO-WAODV routing protocol in high-density scenarios.

[0250] like Fig.13 As shown in the upper middle figure, although the increase in the number of nodes increases the number of network routing opportunities, higher network congestion will cause packet transmission delays. TD-DBESHHO-WAODV reduces congestion and unnecessary route reconstruction time by optimizing the RREQ message forwarding strategy, which reduces the average end-to-end delay.

[0251] like Fig.13 As shown in the figure in the lower middle part, as the number of nodes increases, the number of routing requests generally shows an upward trend. However, due to the judgment of the node topology change degree by the TD-DBESHHO-WAODV routing protocol, the link break frequency of low-speed nodes caused by the movement of high-speed nodes is reduced, and the number of unnecessary routing requests is reduced.

[0252] From the above simulation analysis, it can be concluded that the optimization method based on topology change degree of the present invention effectively reduces the frequency of route discovery and the control overhead of the network in scenarios with high density and fast network topology changes, while improving the overall performance of the network. The TD-DBESHHO-WAODV routing protocol successfully reduces the performance degradation caused by link breakage by avoiding selecting high-speed moving nodes as intermediate routing nodes, ensuring the long-term stable operation of the network. Under the conditions of 50 nodes, 50m / s moving speed and 0s dwell time, compared with the AODV routing protocol, the TD-DBESHHO-WAODV routing protocol has an increase of 12.82% in throughput, and the packet loss rate, average end-to-end delay and number of routing requests have been reduced by 17.08%, 8.78% and 9.22% respectively, showing better adaptability in high-density dynamic scenarios.

[0253] 5. Introduction to the basic principles of Harris Eagle algorithm

[0254] In the Harris Hawk HHO algorithm, the position of the Harris Hawk is taken as a candidate solution, and the best candidate solution of the iteration is the rabbit. In the process of chasing the rabbit, the hawk group adopts different search and siege strategies according to the rabbit's escape energy E and escape probability r. From an iterative perspective, the algorithm has three stages. The first stage is the exploration stage, and the Harris Hawk explores the rabbit within the range. The second stage is the exploration and development conversion stage, and the Harris Hawk discovers the position of the rabbit. The third stage is the development stage. The Harris Hawk uses four strategies to hunt rabbits, namely soft encirclement, hard encirclement, soft encirclement with gradual dive, and hard encirclement with gradual dive. The specific process of the HHO algorithm is as follows: Fig.14 As shown:

[0255] Although the Harris Hawk algorithm has three stages, it is mainly composed of two behaviors: exploration behavior and exploitation behavior. , dynamically choose exploration or exploitation behavior to hunt. The escape energy of a rabbit is defined as:

[0256] (37);

[0257] In the formula, represents the initial escape energy of the rabbit, which is a random number uniformly distributed in the interval (-1,1), t is the current iteration number, and T is the maximum iteration number. Influence, E is generally in a convergent state. Harris Hawk When When it enters the development phase.

[0258] (1) Exploration phase

[0259] At this stage, individuals of the Harris Hawk population randomly inhabit various places, and use their sharp eyes to explore the traces of rabbits within the hunting range. Harris Hawks randomly adopt two exploration strategies to find rabbits. Strategy one is to randomly inhabit and explore within the population range, and strategy two is to migrate and explore based on the location of the population and the location where the rabbit appears. The mathematical expression of its strategy is as follows:

[0260] (38);

[0261] In the formula, and represents the position of the i-th individual in the population among the individuals of the t and t+1 generations, is the position of a random Harris hawk in the current t generation, is the location of the rabbit, q and are all random numbers uniformly distributed in the interval (0,1), i=1,2,3,4, ub and lb are the upper and lower boundaries of the search space respectively, is the average position of all eagles in the eagle group, and the mathematical expression is as follows:

[0262] (39);

[0263] Where N is the number of eagles.

[0264] (2) Development stage

[0265] When a Harris hawk locks onto its target prey, it enters the hunting phase, forming an encirclement around the prey and waiting for an opportunity to launch a surprise attack. However, the actual hunting process is often more complicated. For example, a rabbit that is surrounded may escape, forcing the Harris hawk to constantly adjust its strategy based on the rabbit's escape route. In order to simulate this dynamic hunting behavior, the HHO algorithm designs four strategies to simulate the different response strategies of the Harris hawk during the hunting process. These four strategies are: soft encirclement, hard encirclement, soft encirclement with gradual dive, and hard encirclement with gradual dive. The Harris hawk adjusts its position based on the rabbit's escape energy. Choose the corresponding capture strategy according to the escape probability r, r is a random number randomly distributed in (0,1), Indicates that the probability of escape is small. Indicates high probability of escape.

[0266] (a) Soft encirclement strategy

[0267] when When the rabbit has enough energy, but the probability of escaping is small, the rabbit tries to escape by jumping. The eagle group is not in a hurry to kill the rabbit, but softly surrounds it in the hope of exhausting the rabbit's energy. The specific mathematical expression of this strategy is as follows:

[0268] (40);

[0269] (41);

[0270] In the formula, J is the random jumping intensity of the bunny, is a random number that follows a uniform distribution in the interval (0,1).

[0271] (b) Hard encirclement strategy

[0272] when When the rabbit is about to run out of energy and is basically unable to escape, the rabbit still tries to struggle and escape. At this time, the eagles directly besiege the rabbit and try to catch it. The specific mathematical expression of this strategy is as follows:

[0273] (42);

[0274] (c) Soft encirclement strategy with gradual dive

[0275] when When the rabbit has sufficient physical strength and has a great chance to escape from the encirclement, the eagles try to dive and attack, correct the rabbit's escape direction, and use Levy flight to more tightly surround the rabbit to prevent it from escaping, further consuming the rabbit's physical strength. The specific mathematical expression of this strategy is as follows:

[0276] (43);

[0277] (44);

[0278] (45);

[0279] In the formula, F is the fitness function, dim is the dimension of the problem being solved, S yes dim A random vector of dimension , randomly distributed in the range (0,1). is the expression of the Levy flight function, and its one-dimensional expression is as follows:

[0280] (46);

[0281] In the formula, and is a random number randomly distributed in the range (0,1). The value of is usually 1.5.

[0282] (d) Gradual dive hard encirclement strategy

[0283] when When the rabbit has no energy to escape, it still has a high probability of escaping capture. The eagles try to narrow the encirclement and prevent the rabbit from escaping by Levy flight. When the rabbit is exhausted, they launch a surprise attack on the prey and quickly capture the rabbit. The specific mathematical expression of this strategy is as follows:

[0284] (47);

[0285] (48);

[0286] (49);

[0287] in, , and Z both represent the positions of the population after the eagle group moves.

[0288] The preferred embodiments of the present invention are described in detail above, but the present invention is not limited to the above embodiments.

Claims

1. A method for implementing the AODV routing protocol based on improved HHO and topology change, characterized in that: The AODV routing protocol quantifies the topological change of nodes by evaluating the change of neighbor nodes, the change of signal strength and the survival time of neighbor node links; in high-density areas, high-speed nodes only serve as source nodes or target nodes; in low-density areas, all nodes participate in routing forwarding; when the target node receives all path information, the Harris Eagle algorithm is used to select the optimal path and respond to information based on the optimal path; The method of selecting the optimal path by the Harris Eagle algorithm specifically refers to: after the AODV routing protocol receives all routing requests, the weight is adjusted according to the throughput, packet loss rate and average end-to-end delay, and the routing cost calculation method is optimized in real time; at the same time, the fitness function is used to quantify the expected throughput, packet loss rate and average end-to-end delay to help select a routing path with better performance; The density of the area where the node is located is determined by counting the number of neighbor nodes at a specific time. If the number of neighbor nodes is greater than the threshold, it means that the area where the node is located is a high-density area; otherwise, it is a low-density area. The computation probability p of node i at time t is i,t , the calculation formula is as follows: In the formula, NE i,t represents the number of neighbor nodes of node i at time t, K is a constant; High-speed nodes are determined by calculating the topological change degree TD of the node itself; if the random number generated by the current node is less than the calculated probability p i,t , the node calculates its own topology change degree TD. If the topology change degree TD is greater than the set threshold, the current node is determined to be a high-speed node, and the current node discards the received route request RREQ message, otherwise it forwards the received route request RREQ message; The weights w1, w2 and w3 are the neighbor node change degrees NC i , signal strength variation SV i and link stability TS i The coefficient of topological change degree TD i The calculation formula is as follows: <h2 style=";text-align:left;direction:ltr">TD<h2 style=";text-align:left;direction:ltr"> i <h2 style=";text-align:left;direction:ltr"> =w1·NC<h2 style=";text-align:left;direction:ltr"> i <h2 style=";text-align:left;direction:ltr"> +w2·SV<h2 style=";text-align:left;direction:ltr"> i <h2 style=";text-align:left;direction:ltr"> +w3·(1-TS<h2 style=";text-align:left;direction:ltr"> i <h2 style=";text-align:left;direction:ltr"> ) (2) w1=w2=w3,w1+w2+w3=1 (3); Among them, the neighbor node change degree NC i , signal strength variation SV i and link stability TS i The coefficient of topological change degree TD i The subscript i in each represents node i; The neighbor node change degree is as follows: The formula for the change degree of neighbor nodes of node i within △t time is: In the formula, neighbor i_t is the set of neighbor nodes of node i at time t, neighbor it+△t is the set of neighbor nodes of node i after △t time; when all neighbor nodes of node i are updated to new nodes after △t time, the change degree of neighbor nodes is NC i The value is 1; if after △t time, the neighbor node has no change, the neighbor node change degree NC i The value is 0; The signal strength changes are as follows: The formula for the signal strength change of node i within △t time is: RSSI_N it =neighbour it ∩neighbour it+△t (5) Where, RSSI j (i, t+△t) is the signal strength value of node i’s neighbor node j at time t+△t; RSSI_N i_t It represents the set of nodes that have been neighbors of node i between time t and t+△t. RSSI j (i, t) is the signal strength value of node i’s neighbor node j at time t; when node i loses contact with all its original neighbor nodes after △t time and is replaced by new neighbor nodes, the signal strength change SV i The value is 1; if after △t time, the neighbor node does not change, and the corresponding signal strength does not change, then the signal strength change degree SV i The value is 0; The link stability is as follows: The formula for link stability of node i in time T is: Where T is the link determination period, link(i,j) is the link maintenance time between node i and neighbor node j within the period of T; when node i's link is broken within T, the maintenance time of each link is 0, then the link maintenance time TS i is 0; if all links maintain a stable connection within the T period, and the maintenance time of each link is T, then the link maintenance time TS i is 1.

2. The method for implementing the AODV routing protocol based on improved HHO and topology change degree as claimed in claim 1, characterized in that: The quality of routing is evaluated by the congestion of routing nodes, signal quality, energy consumption speed and number of routing hops.

3. The method for implementing the AODV routing protocol based on improved HHO and topology changeability as claimed in claim 2, characterized in that: The congestion levels are as follows: In time period t, the congestion degree CL of the node is expressed by the following formula: Where C is the capacity of the queue, and L(τ) is the queue length at time point τ; The signal quality is as follows: For two nodes of the same type at a distance x, the signal quality RSSI between them is calculated using the following formula: In the formula, G r is the receiving antenna gain, G t is the transmit antenna gain, P t is the transmission power, f is the frequency used for wireless network communication, and c is the speed of light; According to the signal quality between two nodes, the threshold RSSI_TH of the signal quality is calculated. The formula is as follows: Where R represents the range of the antenna; if the RSSI from a node to its neighbor node is less than the RSSI_TH threshold, all messages from the neighbor node are discarded; The energy consumption rates are as follows: Energy consumption rate E avg The specific calculation formula is as follows: In the formula, E tx is the energy consumed to send each bit of data, D t is the length of the data sent, E rx is the energy consumed per bit of data received, D r is the length of the received data, P i is the power consumption in idle monitoring mode, t i is the time in idle monitoring mode, T avg is the detection cycle; The specific routing hop count is: The number of hops required for each message to be delivered from the source node to the destination node; Route weighted scoring, as follows: For each route, a weighted score RWS is calculated based on its route hop count, signal quality, congestion of intermediate nodes, and energy consumption rate. Specifically, the weighted score RWS is composed of the total hop count Hop of the route, the congestion of the i-th intermediate node C i , the energy consumption rate E of the i-th node i The signal quality RSSI and SQ of each section of the route are weightedly calculated to obtain the result; q1, q2, q3 and q4 are set to be the weights of the number of route hops, signal quality, congestion degree and energy consumption speed respectively, and q1+q2+q3+q4=1.

4. The method for implementing the AODV routing protocol based on improved HHO and topology change degree as claimed in claim 3, characterized in that: The Harris Hawk algorithm adds nonlinear escape energy, chaos mapping, dynamic reverse learning, and the spiral flight mechanism of the vulture search algorithm; The nonlinear escape energy coefficient is as follows: Taking the nonlinear factor as the new escape energy coefficient, the formula is as follows: In the formula, E0 represents the initial escape energy of the rabbit, which is a random number uniformly distributed in the interval (-1,1), iter is the current iteration number, T iter is the maximum number of iterations; The chaos mapping is as follows: x n+1 =αx n (1-x n ) (14) In the formula, α is a constant, x n is the value of the nth iteration, x n+1 is the value of the n+1th iteration; Dynamic reverse learning is as follows: For a solution X in the search space, its reverse solution X O is defined as follows: X o =lb+ub-X (15) In the formula, lb is the lower limit of the search space, and ub is the upper limit of the search space; The dynamic reverse learning strategy DOBL expands the search range of the current solution and its symmetric solution to the entire search space, and the corresponding dynamic reverse number X DO The mathematical definition is as follows: X DO =X+w·rand1(rand2(lb+ub-X)-X) (19) Where w is the dynamically adjusted weight factor, rand i is a random number uniformly distributed between (0,1), i=1,2; The DOBL strategy is introduced to improve the Harris Hawk algorithm: Step 1, calculate the dynamic reverse learning population according to formula (19); Step 2, merge the dynamic reverse learning population with the original population; Step 3, sort the fitness values ​​of the current Hawk population; Step 4, select the first N individuals with better fitness as the new generation population; The fusion vulture search algorithm is as follows: The vulture search algorithm consists of three stages: selecting the search space, searching for prey in the space, and diving to capture the prey; (1) Select the search space The vulture performs random searches within the range, and selects the search area based on the location of the prey and its own experience. The vulture's position update formula is: P i,new =P best +α·rand1·(P mean +P i ) (20) Where P best is the position with the highest fitness among the current vulture individuals, α is the parameter that controls the position change, rand1 is a random number uniformly distributed in the range of (0,1), and P mean is the average position of all vultures, P i is the position of the i-th vulture; (2) Searching for prey in space The vulture selects a spatial position to search for prey and flies in a spiral shape to find the best dive capture position; the formula for updating the flight position using polar coordinates is: P i,new =P i +y(i)·(P i +P i+1 )+x(i)·(P i +P mean ) (21) xr(i)=r(i)·sin(θ(i)),yr(i)=r(i)·cos(θ(i)) (23) θ(i)=α·π·rand2,r(i)=θ(i)+R·rand3 (24) Where P i+1 represents the position of the i+1th vulture before the population is updated; x(i) and y(i) represent the coordinates of the vulture in the plane rectangular coordinate system; r(i) is the polar radius of the polar coordinate; θ(i) represents the polar angle of the polar coordinate; xr(i) and yr(i) represent the horizontal and vertical coordinate components of the vulture in the current polar coordinate system respectively; |xr| and |yr| are the maximum absolute values ​​of the horizontal and vertical coordinates of all solutions, which are used for normalization processing; α and R are parameters that control the spiral search trajectory of the vulture, and the range of variation is (0,5) and (0.5,2) respectively; rand i is a random number uniformly distributed in the range (0,1), i=2,3; (3) Diving to catch prey The vulture will quickly dive towards the target prey from the optimal position in the search space. At the same time, other individuals in the population will also move towards this optimal position and jointly attack; their motion state is described by the polar coordinate equation, and the position update formula is: P i,new =rand4·P best +x1(i)(P i -c1P mean )+y1(i)(P i -c2P best ) (25) xr(i)=r(i)·sinh(θ(i)),yr(i)=r(i)·cosh(θ(i)) (27) θ(i)=α·π·rand5,r(i)=θ(i) (28) Where c1 and c2 represent the intensity of the vulture's dive toward the optimal point and the center point, and rand i is a random number uniformly distributed in the range (0,1), i=4,5; sinh represents the hyperbolic sine function; cosh represents the hyperbolic cosine function; In the Harris Hawk algorithm, the soft encirclement strategy formula of the Harris Hawk is changed, the spiral flight strategy of the vulture search algorithm is introduced, and the formula (21) is used instead, and the formula (24) is modified as follows: In the formula, α is taken as 1.5, t iter is the current iteration number, T iter is the maximum number of iterations, rand i is a random number uniformly distributed in the range (0,1), i=5,6.

5. The method for implementing the AODV routing protocol based on improved HHO and topology changeability as claimed in claim 4, characterized in that: The specific steps of Harris Hawk algorithm are as follows: Step 1: Input the population size N and the maximum number of iterations T iter , the upper and lower boundaries ub and lb of each dimension of the search space; Step 2: Initialize the population X according to formula (14) i , i=1,2,……,N; Step 3: Update the population position according to the dynamic learning mechanism of formula (19); Step 4: Calculate the fitness function and convert the best fitness value X i , denoted by X rabbit ; Step 5: Update the nonlinear escape energy according to formula (13); Step 6: If the escape energy |E|≥1, update the population position according to formula (30); Where, X i (t) and X i (t+1) represents the position of the i-th individual in the population between the t and t+1 generations, X rand (t) is the position of a random Harris Hawk in the current t generation, X rabbit (t) is the position of the rabbit, q and rand i are all random numbers uniformly distributed in the interval (0,1), i = 1, 2, 3, 4, ub and lb are the upper and lower boundaries of the search space respectively, X m (t) is the average position of all hawks in the flock; Step 7: If |E|≥0.5 and the escape probability r≥0.5, update the population position according to formulas (21) and (29); Step 8: If |E|<0.5 and r≥0.5, update the population position according to formula (31); X i (t+1)=X rabbit (t)-E|X rabbit (t)-X i (t)| (31) In the above formula, E is the escape energy coefficient; Step 9: If |E|≥0.5 and r<0.5, update the population position according to formula (32); Where F is the fitness function; Y1 = X rabbit (t)-E|J*X rabbit (t)-X i (t)|; J is the random jumping intensity of the rabbit; Step 10: If |E|<0.5 and r<0.5, update the population position according to formula (33); Where Y2 = X rabbit (t)-E|J·X rabbit (t)-X m (t)|; Y1, Y2 and Z all represent the position of the population after the eagle group has moved; Step 11: If t iter <T iter If satisfied, proceed to step 3; otherwise, the program ends and outputs the optimal solution and the optimal function value.

Images