Edge computing power network routing method and system based on improved ant colony and link discovery

Abstract

The invention relates to an edge computing power network routing method and system based on improved ant colony and link discovery, and the method comprises the steps: controlling a source node, an intermediate node and a destination node to broadcast a Hello message and receive a HelloRSP, and detecting the existence of neighbor nodes in the communication range of the source node, the intermediate node and the destination node and the link state between the neighbor nodes; service quality information is obtained; sending the service quality information to a destination node through a broadcast routing request message, and selecting a candidate path by using an improved ant colony algorithm; an optimal path is selected from the candidate paths, a routing response message is generated, and main and standby routes are established; and controlling the source node to transmit data along the optimal path, and switching the main route and the standby route according to the change of the link operation state. By improving the ant colony algorithm and self-adaptive link discovery, optimizing computing power network routing and preferentially selecting the optimal path during service forwarding, the real-time performance is ensured, and compared with the prior art, the efficiency and the adaptability are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of computing power networks, and in particular, to an edge computing power network routing method and system based on improved ant colony and link discovery. Background Art

[0002] With the development of new network technologies, computing power networks are composed of a combination of cloud resources and edge nodes, interconnecting distributed computing and storage resources, and realizing on-demand and real-time invocation through unified scheduling, improving resource utilization efficiency. This trend is particularly important in edge combat scenarios where combat units are widely distributed and face challenges such as limited bandwidth, node mobility, and changing demands. The traditional AODV routing protocol features on-demand route discovery and dynamic maintenance of the routing table, adapting to frequent changes in the network topology, but has deficiencies in edge tactical scenarios. First, AODV lacks support for quality of service (QoS), mainly focusing on path reachability and hop count, and not fully considering requirements such as delay, bandwidth, and reliability, making it difficult to meet the requirements of diverse data streams. Second, its single-path maintenance mechanism lacks redundancy, is prone to communication interruption due to link disconnection or node failure, and cannot achieve load balancing and efficient resource utilization. In addition, the regular and high-frequency Hello message exchange increases control overhead, exacerbating network congestion in bandwidth-constrained tactical environments.

[0003] The ant colony algorithm is suitable for optimization problems due to its distributed, self-organizing, and global search capabilities. However, the traditional ant colony algorithm performs well in single-objective optimization but is prone to falling into local optima or being unable to balance multiple objectives in multi-objective optimization. In existing optimization schemes that introduce the ant colony algorithm into AODV, multiple quality of service information (such as energy, hop count, congestion degree) is often converted into a single objective function for optimization, requiring fine-tuning of weight factors. If the parameter selection is inappropriate, some objectives are likely to be overemphasized or ignored, affecting the quality and diversity of the solution. In addition, most of these schemes do not consider multi-path routing strategies, ignore the control overhead problem, such as the fixed-frequency Hello messages occupying too much bandwidth, and do not design an adaptive mechanism for the dynamics of edge tactical scenarios. Therefore, the existing technologies have limitations in multi-objective optimization, path redundancy, and overhead control, and are difficult to meet the real-time and quality of service requirements of tactical communication in edge computing power networks. Summary of the Invention

[0004] The purpose of the present invention is to provide an edge computing power network routing method and system based on improved ant colony and link discovery to solve the problems of limitations in existing technologies in multi-objective optimization, path redundancy, and overhead control.

[0005] To achieve one of the above-mentioned invention purposes, an embodiment of the present invention provides an edge computing power network routing method based on improved ant colony and link discovery, and the method includes,

[0006] By controlling the source node, intermediate nodes, and destination node to broadcast Hello messages and receive Hello_RSP, detect the existence of neighbor nodes within the communication ranges of the source node, intermediate nodes, and destination node and the link status with neighbor nodes, and obtain quality of service information;

[0007] Send the quality of service information to the destination node by broadcasting a routing request message, and select candidate paths using an improved ant colony algorithm;

[0008] Select the optimal path from the candidate paths, generate a routing reply message, and establish primary and backup routes;

[0009] Control the source node to transmit data along the optimal path and perform switching of the primary and backup routes according to changes in the link operating status.

[0010] As a further improvement of an embodiment of the present invention, the method further includes designing a quality of service-aware routing message frame format based on the AODV routing protocol; the routing message frame format includes a routing request RREQ_ACO, a routing reply RREP_ACO, a routing table RT_ACO, a Hello message reply packet Hello_RSP, and a local pheromone table;

[0011] Collect node information of the source node, intermediate nodes, and destination node in the edge computing power network, construct a local routing table, and initialize network topology information;

[0012] Initialize the link status of each node, record the quality of service information between each node and its neighbor nodes within the communication range, and initialize the local pheromone table to store link information; the quality of service information includes bandwidth, delay, and cost;

[0013] According to the service requirements of the source node, trigger the broadcast of a routing request message to establish a session with the destination node.

[0014] As a further improvement of an embodiment of the present invention, the method further includes that the "by controlling the source node, intermediate nodes, and destination node to broadcast Hello messages and receive Hello_RSP, detect the existence of neighbor nodes within the communication ranges of the source node, intermediate nodes, and destination node and the link status with neighbor nodes" includes,

[0015] After a node receives a Hello message, broadcast a Hello_RSP for reply;

[0016] After the node that sent the Hello message receives the Hello_RSP, extract the neighbor node IP address field of the Hello_RSP and determine whether it is the same as the IP address of the node that sent the Hello message;

[0017] If they are the same, determine that the node sending the Hello_RSP is a neighbor node of the node sending the Hello message within the communication range; at this time, obtain the operating status of the link between the node sending the Hello message and this neighbor node, as well as the quality of service information of the link.

[0018] The sending frequency of the Hello message Dynamically adjust according to the remaining bandwidth of the link, including

[0019] Estimate the remaining bandwidth of the link by sending TCP probe packets at an adaptive probing interval , the adaptive probing interval Is expressed as:

[0020]

[0021] Where Is an integer set according to the change situation of the network topology. The more frequent the change of the network topology, The smaller it is;

[0022] When ; where Is the maximum bandwidth threshold, Is the maximum sending frequency;

[0023] When ; where Is the minimum bandwidth threshold, Is the minimum sending frequency;

[0024] When , and enter the silent period after continuously preset times being lower than , and resume sending at the frequency of after the silent period ends.

[0025] As a further improvement of an embodiment of the present invention, the method further includes that the selection of candidate paths using the ant colony algorithm includes

[0026] Finding paths by batches of ants for single-objective optimization, including

[0027] The selection probability for the optimization objective of path delay, the formula is:

[0028]

[0029] The selection probability formula for the optimization objective of path cost is:

[0030]

[0031] The selection probability formula for the optimization objective of link average utilization rate is as follows:

[0032]

[0033] Among them, are the transition probabilities for path delay, path cost, and link average utilization rate respectively, is the pheromone value on the current link (i,j), α is the influence factor of pheromone, and β is the delay of the current link the influence factor of the heuristic value, and γ and σ represent the cost and the influence factor of the heuristic value of the link average utilization rate; In the process of ant path finding, the link pheromone value is dynamically adjusted through the pheromone update mechanism to optimize the path selection, and a candidate path is generated based on the adjusted pheromone value;

[0034] Through multi-objective optimization, the destination node selects the Pareto optimal path from the candidate paths, including,

[0035] The multi-objective optimization sorts the candidate paths by using the fast non-dominated sorting and crowding distance calculation algorithms to generate the Pareto optimal path;

[0036] The optimization objectives of the multi-objective optimization are expressed as:

[0037] The optimization objectives of the multi-objective optimization are expressed as:

[0038]

[0039] Among them, and and are the path delay, path cost, and link average utilization rate from the source node s to the destination node d respectively;

[0040] The delay constraint of the multi-objective optimization is expressed as:

[0041]

[0042] Among them, is the maximum delay threshold;

[0043] The data volume constraint of the multi-objective optimization is expressed as:

[0044]

[0045] Among them, is the data transmission volume between any two nodes n1, n2, is the link bandwidth between any two nodes n1, n2.

[0046] As a further improvement of an embodiment of the present invention, the method further includes that the use of the ant colony algorithm to select candidate paths further includes a pheromone update mechanism, specifically including: Prevent ants from aggregating on the same path in the short term through a local update mechanism, and the local update formula is expressed as:

[0047] where is the local pheromone evaporation factor, is the initial pheromone value; If no ants pass through the path, the pheromone decays according to the formula ; where is the global pheromone evaporation factor Strengthen the Pareto optimal path after each iteration through global update, and the global update formula is expressed as:

[0048] where is a variable, is the reward pheromone constant set according to the path quality.

[0049] As a further improvement of an embodiment of the present invention, the method further includes that the establishment of the primary and backup routes includes:

[0050] Based on the strengthened Pareto optimal path, determine the optimal path through the allocation of primary and backup path markers; among them, the path with the marker bit of 00 is used as the primary path and is the optimal path, and the paths with the marker bits of 01 or 11 are used as backup paths;

[0051] The destination node generates a routing reply message according to the path information in the received routing request message, including the source node address, sequence number, and reverse path information, as well as the primary and backup path marker bits of the preferred path, in combination with the address and sequence number of the destination node itself;

[0052] Unicast the routing reply message back to the source node along the reverse path, and the intermediate node updates the local routing table after receiving the routing reply message; the local routing table includes fields such as the destination node address, sequence number, hop count, next hop, and primary and backup path marker bits.

[0053] As a further improvement of an embodiment of the present invention, the method further includes that the switching of the primary and backup routes according to the change of the link operating state includes:

[0054] The source node preferentially selects the primary path for data transmission;

[0055] Automatically switch to the backup path when the primary path fails;

[0056] Monitor the link status by continuously receiving Hello_RSP messages. If the link status does not meet the quality of service requirements or all paths fail, trigger the rebroadcast of the routing request message to optimize the routing.

[0057] As a further improvement of an embodiment of the present invention, the method further includes that the switching of the primary and backup routes according to the change of the link operating state further includes,

[0058] Repeat the processes of broadcasting the routing request message, selecting candidate paths, and generating routing reply messages through a preset number of iterations;

[0059] Update the pheromone value and the primary and backup path flag bits in each iteration to optimize the routing to adapt to dynamic network changes.

[0060] As a further improvement of an embodiment of the present invention, the method further includes that optimizing the message overhead by compressing the routing request message and the Hello_RSP message includes, compressing the quality of service information fields in the routing request message and the Hello_RSP message through differential coding and cumulative level method, specifically including:

[0061] Divide the delay, cost, and utilization rate into multiple levels, and each level is represented by a fixed number of bits;

[0062] Accumulate the level values per hop and store them in the RREQ_ACO message. Discard the message when the cumulative value exceeds a preset threshold;

[0063] Store the quality of service parameter values on the adjacent node links in the Hello_RSP message.

[0064] To achieve one of the above invention purposes, an embodiment of the present invention provides an edge computing power network routing system based on improved ant colony and link discovery. The system includes a source node, an intermediate node, and a destination node;

[0065] The source node, intermediate node, and destination node respectively include: a multi-path routing selection module, a status link adaptive discovery module, a data compression module, and a data sending module;

[0066] The multi-path routing selection module is used to optimize the selection of the AODV routing path based on the improved ant colony algorithm and calculate the transfer probability for a single quality of service optimization target;

[0067] Among them, the purpose of the multi-path routing selection module in the destination node also includes multi-objective optimization of all solutions obtained by single-objective optimization, and using the fast non-dominated sorting combined with the crowding distance calculation algorithm to solve the Pareto optimal solution; when each iteration of the ant colony algorithm is completed, the multi-path routing selection module in the destination node selects a preset number of paths from the Pareto optimal solution for routing reply to form a forward route, and at the same time performs global pheromone update;

[0068] The status link adaptive discovery module is used to monitor the remaining bandwidth capacity of the link, dynamically adjust the sending frequency of Hello messages according to the remaining bandwidth capacity of the link, and reply to the Hello messages of the received neighbor nodes to maintain the service quality parameters of the neighbor link;

[0069] The data compression module is used to compress the routing request packet and the new fields in Hello_RSP to reduce the overhead of control messages;

[0070] The data sending module is used to send all control messages.

[0071] Compared with the prior art, the present invention provides an edge computing power network routing method based on improved ant colony and link discovery. Aiming at the edge tactical scenario, based on the AODV protocol, through the improved ant colony algorithm and adaptive link discovery, the computing power network routing is optimized. The improved ant colony algorithm first conducts single-objective exploration, and then combines the fast non-dominated sorting and the crowding distance calculation algorithm to achieve multi-objective optimization, providing high-quality Pareto optimal solutions, improving path diversity and service quality. The routing selection simultaneously optimizes delay, cost, and link utilization rate, supports multi-path strategies, and enhances redundancy and load balancing. Differential coding is used to compress control messages and the Hello frequency is dynamically adjusted to reduce bandwidth overhead. When forwarding services, the optimal path is preferentially selected to ensure real-time performance, and the efficiency and adaptability are significantly improved compared with the prior art. Brief Description of the Drawings

[0072] Figure 1 is the overall flowchart of the edge computing power network routing method based on improved ant colony and link discovery according to the present invention.

[0073] Figure 2 is the schematic diagram of optimizing AODV routing path finding based on the improved ant colony algorithm of the edge computing power network routing method based on improved ant colony and link discovery according to the present invention.

[0074] Figure 3 is the flowchart of establishing the AODV routing scheme optimized by the improved ant colony algorithm of the edge computing power network routing method based on improved ant colony and link discovery according to the present invention.

[0075] Figure 4It is the overall architecture diagram of the edge computing power network routing method based on improved ant colony and link discovery according to the present invention.

[0076] Figure 5 It is the module architecture diagram of the nodes of the edge computing power network routing system based on improved ant colony and link discovery according to the present invention. Detailed implementation manners

[0077] The present invention will be described in detail below in conjunction with the specific implementation manners shown in the drawings. However, these implementation manners do not limit the present invention, and any structural, method, or functional transformation made by those of ordinary skill in the art based on these implementation manners is included within the protection scope of the present invention.

[0078] The embodiments of the present invention will be described in detail below. The examples of the embodiments are shown in the drawings, where the same or similar reference numerals represent the same or similar elements or elements with the same or similar functions from beginning to end. The embodiments described below by referring to the drawings are exemplary and are only used to explain the present invention and should not be construed as a limitation of the present invention.

[0079] In the first embodiment of the present invention, the present invention provides an edge computing power network routing method based on improved ant colony and link discovery, as Figure 1 shown, the method includes,

[0080] S1: By controlling the source node, intermediate node, and destination node to broadcast Hello messages and receive Hello_RSP, detect the existence of neighbor nodes within the communication range of the source node, intermediate node, and destination node and the link status with the neighbor nodes, and obtain the quality of service information;

[0081] S2: Send the quality of service information to the destination node by broadcasting a routing request message, and use the improved ant colony algorithm to select candidate paths;

[0082] S3: Select the optimal path from the candidate paths, generate a routing reply message and establish the primary and backup routes;

[0083] S4: Control the source node to transmit data along the optimal path and perform the switching of the primary and backup routes according to the change of the link operation status.

[0084] In a specific implementation manner of the present invention, a routing message frame format for quality of service perception is designed based on the AODV routing protocol; the routing message frame format includes a routing request RREQ_ACO, a routing reply RREP_ACO, a routing table RT_ACO, a Hello message reply packet Hello_RSP, and a local pheromone table;

[0085] Collect the node information of source nodes, intermediate nodes, and destination nodes in the edge computing power network, construct a local routing table, and initialize the network topology information;

[0086] Initialize the link state of each node, record the quality of service information between each node and its neighbor nodes within its communication range, and initialize the local pheromone table to store link information; the quality of service information includes bandwidth, delay, and cost;

[0087] According to the service requirements of the source node, trigger the broadcast of a routing request message and establish a session with the destination node.

[0088] Specifically, collect the node information of source nodes, intermediate nodes, and destination nodes in the edge computing power network, including the IP addresses, sequence numbers, and neighbor node identifiers within the communication range of each node, and construct a local routing table RT_ACO by analyzing this information and initialize the network topology information. Then, initialize the link state of each node, use the source node, intermediate node, and destination node as transmitting nodes, detect the existence and link state of neighbor nodes within the communication range of the node by adaptively broadcasting Hello messages and receiving Hello_RSP messages, record the quality of service information with neighbor nodes, and initialize the local pheromone table to store link information; the quality of service information includes link bandwidth, transmission delay, and communication cost, where bandwidth reflects the link transmission capacity, delay represents the data transmission delay from the current node to the neighbor node, and cost measures the link energy consumption and loss; the local pheromone table assigns an initial pheromone value to each neighbor link and stores the corresponding quality of service information. According to the service requirements of the source node, such as establishing a data session with the destination node that meets specific delay or bandwidth requirements, trigger the broadcast of a routing request message RREQ_ACO, and gradually establish a communication session path from the source node to the destination node through the forwarding of neighbor nodes, while updating the local routing tables and pheromone tables of the nodes along the way.

[0089] It should be noted that the frame formats of the routing request RREQ_ACO, routing table RT_ACO, Hello message reply packet Hello_RSP, and local pheromone table proposed by the present invention are shown in Table 1, Table 2, Table 3, and Table 4 respectively.

[0090] Table 1 Routing request RREQ_ACO frame format

[0091]

[0092] Table 2 Routing table RT_ACO format

[0093]

[0094] Table 3 Hello_RSP frame format

[0095]

[0096] Table 4 Local Pheromone Table

[0097]

[0098] In a specific embodiment of the present invention, by controlling the source node, intermediate node, and destination node to broadcast Hello messages and receive Hello_RSP, the presence of neighbor nodes within the communication ranges of the source node, intermediate node, and destination node and the link status with the neighbor nodes are detected. Specifically,

[0099] After a node receives a Hello message, it broadcasts a Hello_RSP for reply;

[0100] After the node that sends the Hello message receives the Hello_RSP, it extracts the neighbor node IP address field of the Hello_RSP and determines whether it is the same as the IP address of the node that sends the Hello message;

[0101] If they are the same, it is determined that the node that sends the Hello_RSP is the neighbor node of the node that sends the Hello message within the communication range; at this time, the operating status of the link between the node that sends the Hello message and this neighbor node and the quality of service information of the link are obtained;

[0102] It should be noted that the source node, intermediate node, and destination node send Hello messages in a broadcast manner. The Hello message contains the IP address and sequence number of the sending node, which are used to identify the node identity and message timeliness. After receiving the Hello message, the neighbor nodes within the communication range generate and broadcast a routing reply message Hello_RSP. The Hello_RSP includes the IP address of this node, the IP address of the neighbor node, and a link information field. The link information field contains the bandwidth, delay, and cost parameters of the link. After receiving the Hello_RSP, the node that sent the Hello message parses the neighbor node IP address field and verifies whether the neighbor node IP address in the Hello_RSP is the same as its own IP address. If they are the same, it is confirmed that the node that sent the Hello_RSP is the neighbor node of the node that sent the Hello message within the communication range. Subsequently, the link information field in the Hello_RSP is extracted to obtain the status of the link between the sending node and this neighbor node, that is, the quality of service information of the link, including bandwidth, delay, and cost. The extracted link quality of service information is saved to the local pheromone table. The local pheromone table records the link pheromone values and corresponding quality of service parameters between each neighbor node and is dynamically updated according to the received Hello_RSP to maintain real-time neighbor relationships and link status. Through the above process, dynamic detection of neighbor nodes within the communication range and continuous monitoring of link status are realized, providing basic data for subsequent route selection.

[0103] Furthermore, the sending frequency of the Hello message is dynamically adjusted according to the remaining bandwidth of the link, including estimating the remaining bandwidth of the link by sending TCP probe packets at an adaptive probe interval

[0104] The adaptive probe interval is expressed as:

[0105]

[0106] where is an integer set according to the change situation of the network topology. The more frequent the network topology changes, the smaller it is;

[0107] When ; where is the maximum bandwidth threshold, is the maximum sending frequency;

[0108] When ; where​​​ is the minimum bandwidth threshold, is the minimum transmission frequency;

[0109] When , , and enter the silent period after being lower than for a continuous preset number of times. After the silent period ends, resume transmission at frequency.

[0110] It should be noted that a state - adaptive link discovery mechanism provided by the present invention optimizes the link discovery problem in a mobile ad - hoc network, aiming to improve network resource utilization and the real - time performance of link discovery. In this mechanism, the transmission frequency of Hello messages is dynamically adjusted according to the remaining link bandwidth capacity to adapt to the change of network bandwidth. When the link bandwidth is low, the detection frequency of Hello messages is reduced to avoid excessive bandwidth occupation and reduce network congestion and packet loss; when the bandwidth is large, Hello messages with a higher detection frequency will not have too much impact on the network burden, but instead help to detect network topology changes more frequently and maintain the real - time performance and accuracy of the routing table.

[0111] Furthermore, to implement this adaptive mechanism, the present invention proposes a dynamic adjustment strategy based on the remaining link bandwidth capacity. This mechanism first sets bandwidth thresholds according to the bandwidth level of the link, namely the low - bandwidth threshold and the high - bandwidth threshold, and these thresholds can be adjusted according to the network environment and requirements. Determine the transmission frequency of Hello messages according to the remaining bandwidth of the current link. The specific rules for the current bandwidth remaining amount are as follows:

[0112] A. When the remaining bandwidth is greater than , the transmission frequency of Hello messages is set to the most frequent frequency (a fixed high frequency), because the link bandwidth is sufficient and nodes can exchange control messages more frequently to ensure the rapid update of network topology and routing information.

[0113] B. When the bandwidth is between and , dynamically adjust the transmission frequency of Hello messages through proportional calculation.

[0114] C. When the link bandwidth is lower than , the frequency will be reduced to the lowest frequency for transmission; if the probe packet continuously detects that the link bandwidth remains lower than for m (such as m = 3) times at the lowest frequency, the node stops sending Hello messages and enters the silent period. The duration of the silent period is preset to a fixed value (such as 2 seconds) or dynamically adjusted according to the network condition. After the silent time ends, continue to transmit at the lowest frequency Send a Hello message to avoid bandwidth waste.

[0115] The network node estimates the remaining bandwidth of the link by periodically sending probe packets . The probe time interval ∆t is adapted to the Hello message sending interval to avoid unnecessary burdens caused by frequent measurements and message transmissions.

[0116] Furthermore, by setting HELLO_INTERVAL to control the sending frequency of each Hello message, as the link bandwidth changes, HELLO_INTERVAL will be dynamically adjusted. The ALLOWED_HELLO_LOSS timer is used to determine whether a Hello message from a certain neighbor has not been received for a long time. If a Hello message from a certain neighbor has not been received within ALLOWED_HELLO_LOSS * HELLO_INTERVAL time, it is considered that the link may be broken and route recovery is triggered. To adapt to the dynamically adjusted HELLO_INTERVAL, the value of ALLOWED_HELLO_LOSS should be dynamically adjusted according to the change of the Hello message sending interval to ensure that the link break can be reasonably judged while the sending interval changes. The maximum number of allowed losses ALLOWED_HELLO_LOSS can be set to a value proportional to HELLO_INTERVAL, usually ALLOWED_HELLO_LOSS = k * HELLO_INTERVAL, where k is a constant coefficient representing the maximum number of tolerated Hello message losses. When the bandwidth is large, ALLOWED_HELLO_LOSS is set to a smaller value (such as 2, that is, tolerating 2 Hello message losses); while when the bandwidth is small, it is set to a larger value (such as 4 or 5). The selection of the coefficient k is recommended to be optimized through experiments and the actual usage of the network. Different network environments may require different tolerances.

[0117] In a specific embodiment of the present invention, the ant colony algorithm is used to select candidate paths. Specifically,

[0118] Ants search for paths in batches for single-objective optimization, including,

[0119] For the selection probability with the optimization objective of path delay, the formula is:

[0120]

[0121] For the selection probability formula with the optimization objective of path cost is:

[0122]

[0123] The selection probability formula for the optimization objective of link average utilization rate is as follows:

[0124]

[0125] Among them, are the transition probabilities for path delay, path cost, and link average utilization rate respectively, is the pheromone value on the current link (i, j), α is the influence factor of pheromone, and β is the delay of the current link the influence factor of the heuristic value, and γ and σ represent the cost and the influence factor of the heuristic value of the link average utilization rate respectively; In the process of ant path finding, the link pheromone value is dynamically adjusted through the pheromone update mechanism to optimize the path selection, and a candidate path is generated based on the adjusted pheromone value;

[0126] It should be noted that the improvement of the existing ant colony algorithm often converts multiple objectives into a single comprehensive objective, making it difficult to effectively balance each optimization objective. Therefore, the present invention proposes an improvement strategy: ants calculate the transition probability in batches with different single objectives during the path finding process, and then perform multi-objective optimization on the single-objective solutions after reaching the destination node. Through the fast non-dominated sorting and crowding distance calculation algorithms, the Pareto optimal solutions are obtained to explore diverse path characteristics, providing a more abundant and representative solution set for multi-objective optimization, and finally obtaining high-quality Pareto front solutions.

[0127] Furthermore, through multi-objective optimization, the destination node selects the Pareto optimal path from the candidate paths. Specifically,

[0128] the multi-objective optimization sorts the candidate paths by using the fast non-dominated sorting and crowding distance calculation algorithms to generate the Pareto optimal path;

[0129] The optimization objectives of the multi-objective optimization are expressed as:

[0130] Among them,

[0131]

[0132] Among them, , , are the path delay, path cost, and link average utilization rate from the source node s to the destination node d respectively;

[0133] The delay constraint of the multi-objective optimization is expressed as:

[0134]

[0135] Among them, is the maximum delay threshold;

[0136] The data volume constraint for multi-objective optimization is expressed as:

[0137]

[0138] where is the data transmission volume between any two nodes n1 and n2, is the link bandwidth between any two nodes n1 and n2.

[0139] It should be noted that

[0140]

[0141]

[0142]

[0143] where is the total number of links on the transmission path has, is the link transmission data volume, is the link bandwidth, is the node transmission energy consumption, is the link loss, is the link delay, is the node processing delay.

[0144] Furthermore, in order to ensure the reachability and stability of the routing, it is also necessary to meet the constraint conditions in terms of delay, link, etc.; the delay constraint means that the maximum delay required to meet the routing timeliness of the edge ad-hoc network needs to be satisfied; the data volume constraint means that the data volume transmitted between any two nodes needs to be less than the link bandwidth between the nodes to ensure the reachability of the routing.

[0145] In a specific embodiment of the present invention, using the ant colony algorithm to select candidate paths further includes a pheromone update mechanism, specifically,

[0146] By the local update mechanism, ants are prevented from aggregating on the same path in the short term, and the local update formula is expressed as:

[0147]

[0148] where is the local pheromone evaporation factor, is the initial pheromone value;

[0149] If no ants pass through the path, the pheromone decays according to the formula ; where is the global pheromone evaporation factor

[0150] Reinforce the Pareto optimal path after each iteration through global update, and the global update formula is expressed as:

[0151]

[0152] Wherein, is a variable, is the reward pheromone constant set according to the path quality.

[0153] It should be noted that during the process of path selection by ants, local update and global update of pheromone are adopted. By introducing local volatilization, the pheromone on the path decays over time, thus preventing the pheromone concentration on a single path from being too high, avoiding falling into local optimum, and enabling ants to continuously explore different path combinations, thereby enhancing the global search ability.

[0154] Furthermore, by introducing a global pheromone volatilization factor, when a certain path has no ants passing by, the global volatilization mechanism can maintain the dynamic balance of the path pheromone concentration, prompting ants to disperse to a wider area for search and enhancing the global search ability. For the global update rule of pheromone, global update is performed after each iteration.

[0155] Furthermore, in the present invention, after each iteration of the ant colony algorithm is completed, at the destination node, M (M = 3) Pareto optimal paths that meet the service quality constraint conditions are selected by using the methods of fast non-dominated sorting and crowding distance calculation for global pheromone update. Through the action of the local and global pheromone update mechanisms during the iterative process, ants gradually concentrate on the optimal paths that meet the service quality constraints.

[0156] Preferably, based on the traditional AODV protocol, the present invention uses an improved ant colony algorithm to propose a dynamic adaptive reactive service quality priority routing mechanism. The routing request (RREQ_ACO) carries QoS information such as path delay and cost. Nodes dynamically evaluate the service quality during path selection and adaptively update the pheromone concentration according to the link state and QoS requirements. The pheromone of the paths that meet the QoS and perform excellently is enhanced, attracting more ants to choose; if the link deteriorates or a better path appears, the pheromone decays, and the path selection biases towards the new path. This strategy improves the adaptability of the routing to the link requirements in a highly dynamic environment.

[0157] Furthermore, the schematic diagram of optimizing the AODV routing based on the improved ant colony algorithm is as Figure 2 shown. Figure 2In it, in each iteration of the routing algorithm, the forward ants (RREQ_ACO) first calculate the transition probabilities in batches based on a single objective, and start from the source node to find a path that meets the QoS constraints. When all the ants exceed the maximum survival time, the destination node performs multi-objective optimization on the obtained paths, and uses the fast non-dominated sorting and crowding distance calculation algorithm to select M optimal solutions.

[0158] Furthermore, in the routing establishment mechanism, considering that the target scenario is a distributed environment, this routing establishment mechanism also adopts a fully distributed manner. To ensure that all ants from the source node in one iteration have reached the destination node or been discarded, after receiving the first ant, the destination node waits for the maximum survival time of the ants before starting the multi-objective optimization.

[0159] In a specific embodiment of the present invention, a primary and backup route is established. Specifically,

[0160] Based on the enhanced Pareto-optimal path, the optimal path is determined by allocating the primary and backup path flag bits; among them, the path with the flag bit of 00 is used as the primary path and is the optimal path, and the paths with the flag bits of 01 or 11 are used as backup paths;

[0161] The destination node generates a routing reply message according to the path information in the received routing request message, including the source node address, sequence number, and reverse path information, as well as the primary and backup path flag bits of the preferred path, in combination with the address and sequence number of the destination node itself;

[0162] The routing reply message is unicast back to the source node along the reverse path. After receiving the routing reply message, the intermediate node updates the local routing table; the local routing table includes fields such as the destination node address, sequence number, hop count, next hop, and primary and backup path flag bits.

[0163] It should be noted that after the destination node receives all the forward ants (i.e., the routing request message RREQ_ACO), it determines the optimal path based on the enhanced Pareto-optimal path by allocating the primary and backup path flag bits. Specifically, the destination node uses the fast non-dominated sorting combined with the crowding distance calculation algorithm to screen out M Pareto-optimal paths (M is usually set to 3) that meet the quality of service constraints from the candidate paths, and assigns the primary and backup path flag bits to each path. Among them, the path with the flag bit of 00 is designated as the primary path, which is the optimal path and is preferentially used for data transmission; the paths with the flag bits of 01 or 11 are respectively designated as the first backup path and the second backup path, serving as alternative choices when the primary path fails. The destination node constructs the routing reply message RREP_ACO according to the path information in the received routing request message RREQ_ACO, as well as the primary and backup path flag bits of the preferred path, in combination with the address and sequence number of the destination node itself. The frame structure of the RREP_ACO is shown in Table 5. The routing reply message RREP_ACO is unicast back to the source node along the reverse path. After receiving the RREP_ACO, the intermediate nodes along the way update the local routing table RT_ACO according to the information in the message. The local routing table includes the destination node address, the destination sequence number (used to judge the freshness of the route), the hop count (recording the distance to the destination node), the next-hop address (pointing to the node that sends the RREP_ACO), the primary and backup path flag bit field (identifying the path priority), and the time to live (ensuring the validity of the route), thus completing the establishment of the forward route and providing support for subsequent data transmission.

[0164] Table 5 Frame Format of the Routing Reply Message RREP_ACO

[0165]

[0166] In a specific embodiment of the present invention, the switching of the primary and backup routes according to the change of the link operating state includes

[0167] The source node preferentially selects the primary path for data transmission;

[0168] Automatically switches to the backup path when the primary path fails;

[0169] Monitors the link state by continuously receiving the Hello_RSP message. If the link state does not meet the quality of service requirements or all paths fail, triggers the rebroadcast of the routing request message to optimize the route.

[0170] Repeats the process of broadcasting the routing request message, selecting the candidate paths, and generating the routing reply message through a preset number of iterations;

[0171] Updates the pheromone value and the primary and backup path flag bits in each iteration to optimize the route to adapt to the dynamic network changes.

[0172] It should be noted that the source node preferentially selects the primary path with the primary and backup path flag bits being 00 in the routing table RT_ACO for data transmission to ensure the optimal quality of service. If the primary path fails due to node movement, link interruption, or QoS exceeding the threshold, it automatically switches to the backup path, preferentially selecting the path with the flag bit being 01, and then selecting the path with the flag bit being 11 when the former is unavailable. By continuously receiving Hello_RSP messages sent by neighbor nodes at each node, the link status is monitored in real time. If it is detected that the link status does not meet the quality of service requirements, such as the path delay exceeding the threshold or the bandwidth being insufficient, the source node re-broadcasts RREQ_ACO to initiate path discovery. If all paths (primary path and backup path) fail, the source node also re-broadcasts RREQ_ACO to restore the connection with the destination node through new path exploration. Through iterative optimization of the routing, including repeating RREQ_ACO broadcast, candidate path selection, and RREP_ACO generation through a preset maximum number of iterations. In each iteration, the intermediate node updates the pheromone value in the local pheromone table according to the transition probability, and adopts local or global update rules to strengthen the preferred path. At the same time, the destination node reallocates the primary and backup path flag bits according to the Pareto optimal solution, and the flag bits are dynamically adjusted according to the optimization degree of path delay, cost, and link utilization, so as to improve the adaptability of the routing to network dynamic changes and service demand fluctuations, and ensure the optimal path and stable QoS.

[0173] In a specific embodiment of the present invention, optimizing the message overhead by compressing the routing request message and Hello_RSP message includes compressing the quality of service information fields in the routing request message and Hello_RSP message through differential coding and cumulative level method, specifically including:

[0174] Dividing the delay, cost, and utilization into multiple levels, and each level is represented by a fixed number of bits;

[0175] Accumulating the hop-by-hop level values and storing them in the RREQ_ACO message, and discarding the message when the cumulative value exceeds the preset threshold;

[0176] Storing the quality of service parameter values on the link between adjacent nodes in the Hello_RSP message.

[0177] It should be noted that in order to control the overhead of the routing request, the method of differential coding plus cumulative level is used to compress the three fields of the newly added path delay, path cost, and path average link utilization in RREQ_ACO. The path delay, cost, and average link utilization are calculated by recording the level differences between each hop and accumulating them.

[0178] In a specific implementation scenario of the present invention, link delay, cost, and average link utilization are divided into multiple levels (for example, 4 levels, represented by 2 bits to indicate differences), and each level represents a range. For example, the delay range for each level is defined as

[0179] Level 0 (00): The delay is very low (<10ms)

[0180] Level 1 (01): The delay is relatively low (10ms~50ms)

[0181] Level 2 (10): The delay is medium (50ms~100ms)

[0182] Level 3 (11): The delay is relatively high (>100ms)

[0183] Before RREQ_ACO is sent, the path delay, cost, and average link utilization fields are initialized to 0. For each hop passed, the current node adds the level value of this hop to the corresponding cumulative field in RREQ_ACO. This cumulative field can be represented by 4 bits (it can also be 3, 5, etc.), and the number of bits depends on the set level threshold.

[0184] When RREQ_ACO reaches the destination node, the cumulative delay, cost, and link utilization of the entire path can be obtained for calculating the estimated value of the objective function. If the cumulative level value reaches the set threshold (for example, the maximum for 4 bits is 15), then RREQ_ACO is discarded, which is equivalent to setting a maximum tolerance for the quality of service of the path, and exceeding the limit is considered not meeting the QoS requirements.

[0185] In a specific implementation scenario of the present invention, as Figure 3 、 4 shown, when a session needs to be established between the source node and the destination node, and there is no path to the destination node that meets the session quality of service requirements in the current routing table, reactive routing establishment starts. The routing establishment steps are as follows:

[0186] Step 1: In the initial stage, first initialize the neighbor node list of each node. Nodes within the communication range are included in the local node's neighbor list, and the IP addresses of neighbor nodes are saved. At the same time, neighbor nodes that meet the quality of service requirements are stored in the routing table. In the initialization stage, an initial pheromone value is set for the links within the communication range, and the pheromone update starts. After initialization is completed, the source node broadcasts RREQ_ACO to all neighbor nodes, and at the same time updates the local pheromone table. The pheromone value of the link between the source node and neighbor nodes is updated according to the local update rule of pheromone, and the rest are updated according to the pheromone decay update formula. Proceed to Step 2;

[0187] Step 2: When a neighbor node (intermediate node) receives RREQ_ACO, it first determines whether it has processed this message according to the source node address, sequence number, and taboo table in RREQ_ACO. If so, it directly discards the message to avoid loops; otherwise, it proceeds to Step 3.

[0188] Step 3: Node j checks whether it is the destination node. If it is, it proceeds to Step 5; otherwise, it extracts the stored quality-of-service related values from RREQ_ACO, combines them with its own node processing delay and average link utilization constraints, and judges the path quality-of-service constraint conditions. If the conditions are not met, it discards RREQ_ACO; otherwise, it adds the sequence number and source node address carried by RREQ_ACO to the local identification list, and performs addition or update operations on the local routing table and update of the local pheromone table for the path from the source node to the current node.

[0189] Step 4: The relay node calculates the transfer probability of forwarding from the previous hop (assumed to be i) to the current node j according to the transfer probability normalization factor carried in the RREQ_ACO packet and the quality-of-service information stored in the local pheromone table. , and the current node forwards the request packet with the transfer probability size.

[0190] All relay nodes forwarding RREQ_ACO update the routing table according to the following Rule A.

[0191] A: The relay node forwarding RREQ_ACO needs to write some information of this RREQ_ACO message into the routing table and create a new entry to establish a reverse path to the destination node. Specifically, the following entries need to be updated:

[0192] a. Destination IP address: Add the source IP address in RREQ_ACO to the routing table to record the routing entry for the destination node.

[0193] b. Destination sequence number: Write the source node sequence number in RREQ_ACO into the routing table.

[0194] c. Hop count: Increment the hop count in RREQ_ACO by 1 and record it in the routing table.

[0195] d. Next hop: Record the sending node (i.e., the previous hop node) of this RREQ_ACO message as the next hop node to the destination node. This is an important step in establishing the reverse path.

[0196] e. Upstream node list: Record the information of the previous hop node in the upstream node list. The upstream node list is used to notify these nodes for route repair when the link breaks.

[0197] f. Survival Time: Set the survival time for this routing entry (node system time + ACTIVE_ROUTE_TIMEOUT), which indicates the validity period of this route. When the survival time expires, if there is no new update, the route will be removed.

[0198] Step 5: When the RREQ_ACO of each forward ant sent from the source node reaches the destination node from the source node, that is, when all ants complete one iteration, the destination node calculates the Pareto front for all solutions using fast non-dominated sorting based on all the single-objective optimized paths (i.e., solutions) obtained, and then calculates M (M = 3) Pareto-optimal solutions in combination with the crowding distance to achieve multi-objective optimization. The closer the solution (path) is to the Pareto front and the farther the crowding distance from the adjacent solution, the smaller the value of the primary and backup path flag bits. For these M forward ants, extract information from them, construct the corresponding RREP_ACO, and at the same time destroy these forward ants. The RREP_ACO reaches the source node along the corresponding reverse path. During this process, global pheromone update is performed on this path, and at the same time, the intermediate nodes update the routing table.

[0199] B: Routing table update rule when RREP_ACO returns:

[0200] a. Destination IP address: Add the destination IP address in the RREP_ACO to the routing table.

[0201] b. Destination node serial number: Update the destination node serial number in the RREP_ACO to the routing table.

[0202] c. Hop count: Increment the hop count value in the RREP_ACO by 1 and update it to the routing table.

[0203] d. Next hop: Record the previous hop node address of the RREP_ACO as the next hop to reach the destination node.

[0204] e. Survival Time: Write the survival time in the RREP_ACO to the routing table. Regularly check whether it has expired, and the expired routes will be deleted.

[0205] f. Primary and backup path flag bits: Put the primary and backup path flag bits in the RREP_ACO into the optimization objective value field of the routing table.

[0206] Step 6: Repeat Steps 1 - 4 until the maximum number of iterations is reached. Then, proceed to Step 7.

[0207] Step 7: For these M paths, construct the corresponding RREP_ACO. Finally, it will unicast back to the source node along the path of the RREQ_ACO. Each time it passes through an intermediate node, the node updates the local routing table information.

[0208] Step 8: The RREP_ACO reaches the source node, and the ant is discarded, thus establishing a forward route from the source node to the destination node.

[0209] Step 9: After establishing a path to the destination node, the source node sends data, and the data packet is forwarded according to the information in the routing table. If there are multiple paths from a node to the destination node, the path with the smallest primary / backup path flag in the routing table is selected. If the optimal path fails during data transmission, the backup path is selected.

[0210] In the second embodiment of the present invention, a routing system for an edge computing power network based on improved ant colony and link discovery is provided. As Figure 4 、 5 shown, it includes a source node, intermediate nodes, and a destination node;

[0211] The source node, intermediate nodes, and destination node respectively include: a multi-path routing selection module, a status link adaptive discovery module, a data compression module, and a data sending module;

[0212] The multi-path routing selection module is used to optimize the selection of the AODV routing path based on the improved ant colony algorithm and calculate the transfer probability for a single quality of service optimization objective;

[0213] Among them, the uses of the multi-path routing selection module in the destination node also include performing multi-objective optimization on the solutions obtained from all single-objective optimizations, and using the fast non-dominated sorting combined with the crowding distance calculation algorithm to solve the Pareto optimal solution; when each iteration of the ant colony algorithm is completed, the multi-path routing selection module in the destination node selects a preset number of paths from the Pareto optimal solution for routing reply to form a forward route, and at the same time performs global pheromone update;

[0214] Further, the multi-path routing selection module of the source node further includes a service message processing module; the service message processing module is used to select the optimal primary path for forwarding service messages and backup paths when service messages arrive.

[0215] The status link adaptive discovery module is used to monitor the remaining bandwidth capacity of the link, dynamically adjust the sending frequency of Hello messages according to the remaining bandwidth capacity of the link, and reply to the Hello messages of the received neighbor nodes to maintain the quality of service parameters of the neighbor link;

[0216] The data compression module is used to compress the routing request packet and the new fields in Hello_RSP to reduce the overhead of control messages;

[0217] The data sending module is used to send all control messages.

[0218] In summary, a routing method for an edge computing power network based on improved ant colony and link discovery provided by the present invention aims at edge tactical scenarios. Based on the AODV protocol, the computing power network routing is optimized by improving the ant colony algorithm and adaptive link discovery. The improved ant colony algorithm first conducts single-object exploration, and then combines the fast non-dominated sorting and crowding distance algorithms to achieve multi-objective optimization, providing high-quality Pareto optimal solutions. The routing selection simultaneously optimizes delay, cost, and link utilization rate, and supports multi-path strategies. Differential coding is used to compress control messages and the Hello frequency is dynamically adjusted to reduce bandwidth overhead. When forwarding services, the optimal path is preferentially selected, significantly improving efficiency and adaptability compared with the prior art.

[0219] It should be understood that although this specification is described according to embodiments, not every embodiment only contains an independent technical solution. This narrative manner of the specification is only for clarity. Those skilled in the art should regard the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

[0220] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the above-described modules can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0221] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed to multiple network modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0222] In addition, in each embodiment of the present application, the functional modules can be integrated into one processing module, or each module can exist physically alone, or two or more modules can be integrated into one module. The above-integrated modules can be implemented in the form of hardware or in the form of a combination of hardware and software functional modules.

[0223] The integrated module implemented in the form of software function modules can be stored in a computer-readable storage medium. The above software function modules are stored in a storage medium and include several instructions for causing a computer system (which may be a personal computer, a server, or a network system, etc.) or a processor to execute some steps of the methods described in various embodiments of the present application. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs that can store program codes.

[0224] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of various embodiments of the present application.

Claims

1. An edge computing power network routing method based on improved ant colony and link discovery, characterized in that: including By controlling the source node, intermediate nodes, and destination node to broadcast Hello messages and receive Hello_RSP, detecting the existence of neighbor nodes within the communication ranges of the source node, intermediate nodes, and destination node and the link status with neighbor nodes, and obtaining quality of service information; Sending the quality of service information to the destination node by broadcasting a routing request message, and selecting candidate paths using an improved ant colony algorithm; Selecting the optimal path from the candidate paths, generating a routing reply message, and establishing primary and backup routes; Controlling the source node to transmit data along the optimal path and switching the primary and backup routes according to changes in the link operating status.

2. The edge computing power network routing method based on improved ant colony and link discovery according to claim 1, characterized in that: It also includes Designing a quality of service-aware routing message frame format based on the AODV routing protocol; the routing message frame format includes a routing request RREQ_ACO, a routing reply RREP_ACO, a routing table RT_ACO, a Hello message reply packet Hello_RSP, and a local pheromone table; Collecting node information of the source node, intermediate nodes, and destination node in the edge computing power network, constructing a local routing table, and initializing network topology information; Initializing the link status of each node, recording the quality of service information between each node and its neighbor nodes within the communication range, and initializing the local pheromone table to store link information; the quality of service information includes bandwidth, delay, and cost; According to the service requirements of the source node, triggering the broadcast of a routing request message and establishing a session with the destination node.

3. The edge computing power network routing method based on improved ant colony and link discovery according to claim 1, wherein: The "By controlling the source node, intermediate nodes, and destination node to broadcast Hello messages and receive Hello_RSP, detecting the existence of neighbor nodes within the communication ranges of the source node, intermediate nodes, and destination node and the link status with neighbor nodes" includes After a node receives a Hello message, broadcasting a Hello_RSP for reply; After the node that sent the Hello message receives the Hello_RSP, extracting the neighbor node IP address field of the Hello_RSP and determining whether it is the same as the IP address of the node that sent the Hello message; If they are the same, determining that the node that sent the Hello_RSP is a neighbor node of the node that sent the Hello message within the communication range; at this time, obtaining the operating status of the link between the node that sent the Hello message and this neighbor node and the quality of service information of the link; The sending frequency of the Hello message According to the remaining link bandwidth Perform dynamic adjustment, including Estimating the remaining link bandwidth by sending TCP probe packets at an adaptive probing interval , the adaptive probing interval is expressed as: Among them, is an integer set according to the change of network topology. The more frequent the change of network topology is, the smaller it is; When then ; wherein is the maximum bandwidth threshold, is the maximum transmission frequency; When then ; wherein is the minimum bandwidth threshold is the minimum transmission frequency; When then and enter the silent period after the continuous preset number of times is lower than After the silent period ends, resume sending at frequency 4. The edge computing power network routing method based on improved ant colony and link discovery according to claim 1, characterized in that: The "selecting candidate paths using the ant colony algorithm" includes Finding paths through ants in batches for single-objective optimization, including For the selection probability with the optimization objective of path delay, the formula is: For the selection probability formula with the optimization objective of path cost is: For the selection probability formula with the optimization objective of link average utilization rate is: Among them, are the transition probabilities for path delay, path cost, and link average utilization respectively, is the pheromone value on the current link (i, j), α is the influence factor of pheromone, and β is the current link delay influence factor of the heuristic value, and γ and σ represent the cost and the influence factor of the link average utilization heuristic value; During the ant path finding process, dynamically adjusting the link pheromone value through a pheromone update mechanism to optimize path selection, and generating candidate paths based on the adjusted pheromone value; Through multi-objective optimization, the destination node selects a Pareto optimal path from the candidate paths, including The multi-objective optimization sorts the candidate paths by using a fast non-dominated sorting and crowding distance calculation algorithm to generate a Pareto optimal path; The optimization objectives of the multi-objective optimization are expressed as: Among them, , , are the path delay, path cost, and average link utilization from the source node s to the destination node d, respectively; The delay constraint of the multi-objective optimization is expressed as: Among them, is the maximum delay threshold; The data volume constraint of the multi-objective optimization is expressed as: Among them, is the data transfer volume between any two nodes n1 and n2, is the link bandwidth between any two nodes n1 and n2.

5. The edge computing power network routing method based on improved ant colony and link discovery according to claim 4, characterized in that: The selection of candidate paths using the ant colony algorithm also includes a pheromone update mechanism, specifically including, Prevent ants from aggregating on the same path in the short term through the local update mechanism. The local update formula is expressed as: Among them, is the local pheromone volatilization factor, is the initial pheromone value; If no ants pass through the path, the pheromone decays according to the formula ; where is the global pheromone evaporation factor Reinforce the Pareto optimal path after each iteration through global update. The global update formula is expressed as: Among them, is a variable, is a reward pheromone constant set according to the path quality.

6. The edge computing power network routing method based on improved ant colony and link discovery according to claim 5, characterized in that: The establishment of the primary and backup routes includes, Based on the reinforced Pareto optimal path, determine the optimal path through the allocation of primary and backup path flag bits; among them, the path with the flag bit of 00 is used as the primary path and is the optimal path, and the paths with the flag bits of 01 or 11 are used as backup paths; The destination node generates a routing reply message according to the path information in the received routing request message, including the source node address, sequence number, and reverse path information, as well as the primary and backup path flag bits of the preferred path, in combination with the address and sequence number of the destination node itself; Unicast the routing reply message back to the source node along the reverse path. After receiving the routing reply message, the intermediate node updates the local routing table; the local routing table includes fields such as the destination node address, sequence number, hop count, next hop, and primary and backup path flag bits.

7. The edge computing power network routing method based on improved ant colony and link discovery according to claim 6, characterized in that: The switching of the primary and backup routes according to the change of the link operating state includes, The source node preferentially selects the primary path for data transmission; Automatically switch to the backup path when the primary path fails; Monitor the link state by continuously receiving Hello_RSP messages. If the link state does not meet the quality of service requirements or all paths fail, trigger the rebroadcast of the routing request message to optimize the routing.

8. The edge computing power network routing method based on improved ant colony and link discovery according to claim 7, characterized in that: The switching of the primary and backup routes according to the change of the link operating state also includes, Repeat the processes of broadcasting the routing request message, selecting candidate paths, and generating the routing reply message through a preset number of iterations; Update the pheromone value and the primary and backup path flag bits in each iteration to optimize the routing to adapt to dynamic network changes.

9. The edge computing power network routing method based on improved ant colony and link discovery according to claim 1, characterized in that: Also includes, Optimizing the message overhead by compressing the routing request message and Hello_RSP message includes compressing the quality of service information fields in the routing request message and Hello_RSP message through differential coding and cumulative level methods, specifically including: Divide the delay, cost, and utilization rate into multiple levels, and each level is represented by a fixed number of bits; Accumulate the level values per hop and store them in the RREQ_ACO message. Discard the message when the cumulative value exceeds the preset threshold; Store the quality of service parameter values on the adjacent node links in the Hello_RSP message.

10. An edge computing power network routing system based on improved ant colony and link discovery, characterized in that: Includes a source node, intermediate nodes, and a destination node; The source node, intermediate nodes, and destination node respectively include: a multi-path routing selection module, a state link adaptive discovery module, a data compression module, and a data sending module; The multi-path routing selection module is used to optimize the selection of the AODV routing path based on the improved ant colony algorithm and calculate the transition probability for a single quality of service optimization objective; Among them, the purpose of the multi-path routing selection module in the destination node also includes multi-objective optimization of all solutions obtained by single-objective optimization, and using the fast non-dominated sorting combined with the crowding distance calculation algorithm to solve the Pareto optimal solution; when each iteration of the ant colony algorithm is completed, the multi-path routing selection module in the destination node selects a preset number of paths from the Pareto optimal solution for routing reply to form a forward route, and at the same time performs global pheromone update; The status link adaptive discovery module is used to monitor the remaining bandwidth capacity of the link, and dynamically adjust the sending frequency of Hello messages according to the remaining bandwidth capacity of the link. At the same time, it replies to the received Hello_RSP messages of neighbor nodes to maintain the quality of service parameters of neighbor links; The data compression module is used to compress the routing request packet and the new fields in Hello_RSP to reduce the overhead of control messages; The data sending module is used to send all control messages.

Images