A method for entanglement request scheduling based on quantum network

Abstract

The present disclosure relates to a method for entanglement request scheduling based on quantum network. The method comprises the following steps: first, establishing a single long-distance request routing path selection method, selecting the end-to-end fidelity highest routing path for each request; then establishing a two-stage request scheduling strategy, and finally establishing a comprehensive evaluation mechanism based on network throughput, fairness index and request service rate. Through the above settings, in the quantum network purification operation scene, the end-to-end fidelity is taken as the key indicator for request priority ranking, a refined fidelity evaluation model is established, the resource demand and application value of different entanglement requests are accurately quantified, the high fidelity requests are preferentially scheduled into the resource allocation queue, the resource loss characteristics in the purification process are dynamically adapted, the dual optimal configuration of entanglement resources in time dimension and space dimension is realized, the quality of service of high value requests is guaranteed, resource idling and waste are avoided, and the overall resource utilization is significantly improved.

Description

Technical Field

[0001] This invention relates to the technical fields of network protocols and quantum technology, and in particular to a method for entanglement request scheduling based on quantum networks. Background Technology

[0002] Quantum networks use quantum states as information carriers and rely on core principles such as quantum entanglement and quantum teleportation to interconnect distributed quantum nodes, constructing a revolutionary new generation of information processing systems. Based on the no-cloning theorem, quantum networks can provide absolutely secure communication in the sense of information theory; at the same time, by sharing entangled resources, they can achieve distributed quantum computing and quantum sensing that break through the limits of classical technology.

[0003] In practical applications of quantum networks, establishing quantum entanglement between two terminal nodes is a crucial prerequisite for various quantum applications. Due to transmission distance limitations, terminal nodes often cannot directly establish entanglement and must rely on relay nodes to perform entanglement swapping operations to achieve indirect connections. However, the entanglement distribution process typically occurs in noisy environments, making it difficult to guarantee the quality (i.e., entanglement fidelity) of end-to-end entangled connections. Since various innovative quantum applications have extremely high requirements for entanglement fidelity, researchers need to improve fidelity through techniques such as entanglement purification; however, purification and swapping operations are probabilistic, and the purification process consumes a large amount of entanglement resources, resulting in a persistent scarcity of entanglement resources in quantum networks. Furthermore, quantum requests in the network often exhibit concurrency and competition, making it difficult for limited entanglement resources to meet all requests immediately.

[0004] In the aforementioned technical solutions, determining the execution order of concurrent requests and formulating a reasonable resource scheduling scheme in scenarios with limited network resources has become a key challenge in the field of quantum networks. This problem is commonly referred to as the entanglement request scheduling problem. Although existing research has proposed some solutions to this problem, none of these solutions fully consider the additional resource consumption caused by purification operations.

[0005] Therefore, it is necessary to provide a new technical solution to improve one or more of the problems existing in the above solutions. Summary of the Invention

[0006] The purpose of this disclosure is to provide a method for scheduling entangled requests based on quantum networks, which achieves optimal utilization of entangled resources by prioritizing the processing of requests with higher end-to-end fidelity.

[0007] A method for entanglement request scheduling based on quantum networks according to the present invention includes the following steps:

[0008] Based on the establishment of entangled connections between adjacent quantum nodes, a routing path selection method for a single long-distance request is established to select the routing path with the highest end-to-end fidelity for each request.

[0009] A two-stage request scheduling strategy is established, which includes a resource pre-allocation stage and a resource replenishment stage.

[0010] During the resource pre-allocation phase, requests within the current request round are executed in descending order of end-to-end fidelity.

[0011] For requests whose requirements are not met in the resource pre-allocation phase, the process proceeds to the resource replenishment phase, where a suboptimal route is selected for execution.

[0012] Establish a comprehensive evaluation mechanism based on network throughput, fairness index, and request service rate.

[0013] Preferably, during the resource pre-allocation phase, the end-to-end fidelity of all requests within the current request round is determined;

[0014] The request with the highest end-to-end fidelity is executed first, and the remaining requests are scheduled and executed in descending order of end-to-end fidelity.

[0015] Preferably, during the resource replenishment phase, requests whose needs were not met during the resource pre-allocation phase are combined with newly arriving request rounds to form a new request set, and the unmet requests are executed with priority.

[0016] Preferably, the resource replenishment phase is executed only once for each of the unmet needs.

[0017] Preferably, when there is more than one unmet request, the execution order of the multiple unmet requests in the resource replenishment phase is consistent with that in the resource pre-allocation phase.

[0018] Preferably, the requests in the newly arriving request round are scheduled and executed sequentially in descending order of end-to-end fidelity.

[0019] Preferably, when the final request round of the resource replenishment phase is completed, if a new unmet request arises, the resource replenishment phase is restarted for the new unmet request.

[0020] Preferably, in the comprehensive evaluation mechanism, for each individual request, the throughput is defined as the number of end-to-end entangled connections established by that request, wherein the throughput is determined by the link capacity of the selected routing path when the selected routing path does not require purification operations.

[0021] Preferably, in the comprehensive evaluation mechanism, the fairness index is Jain's index.

[0022] Preferably, in the comprehensive evaluation mechanism, the request service rate is defined as:

[0023] ,

[0024] in, For the requested service rate, Let r be an element in R, representing the throughput achieved by this request on the selected path. This represents the actual resource requirements of request r. This refers to the number of concurrent source-destination pairs, which is the total number of requests entangled in a set of requests.

[0025] The technical solutions provided by the embodiments of this disclosure may include the following beneficial effects:

[0026] 1. In the quantum network purification operation scenario, this technical solution uses end-to-end fidelity as the key indicator for request priority ranking. By establishing a refined fidelity evaluation model, it accurately quantifies the resource requirements and application value of different entanglement requests, prioritizes high-fidelity requests to enter the resource allocation queue, and dynamically adapts to the resource loss characteristics during the purification process. This achieves optimal allocation of entanglement resources in both time and space dimensions, ensuring the service quality of high-value requests while avoiding resource idleness and waste, and significantly improving the overall resource utilization rate.

[0027] 2. The request scheduling strategy designed in this technical solution effectively alleviates the competition pressure of concurrent requests by adding additional attempts at suboptimal paths;

[0028] 3. Compared to the design flaws of traditional entangled request scheduling algorithms that only focus on fairness or a single performance indicator, this algorithm, through the supplementation of a priority calibration mechanism and suboptimal paths, achieves a comprehensive improvement in network performance while ensuring that all types of requests receive reasonable service opportunities. That is, the scheduling strategy significantly improves the throughput and request service rate of quantum networks, while reducing the average waiting latency of requests. It exhibits better adaptability and practicality in high-concurrency and high-dynamic quantum network environments, providing key technical support for the large-scale deployment and efficient operation of quantum networks. Attached Figure Description

[0029] Figure 1 A flowchart illustrating a method for entanglement request scheduling based on a quantum network in an exemplary embodiment of this disclosure is shown.

[0030] Figure 2 This diagram illustrates a flowchart of requesting a scheduling policy in an exemplary embodiment of this disclosure;

[0031] Figure 3 This diagram illustrates an example of entanglement request scheduling in an exemplary embodiment of this disclosure.

[0032] Figure 4 This diagram illustrates a comparison of throughput at different fidelity thresholds in an exemplary embodiment of this disclosure.

[0033] Figure 5 This diagram illustrates a throughput comparison at different entanglement swap success rates in an exemplary embodiment of this disclosure.

[0034] Figure 6 This diagram illustrates a comparison of throughput at different numbers of SD pairs in an exemplary embodiment of this disclosure.

[0035] Figure 7 This diagram illustrates a comparison of fairness indices at different fidelity thresholds in an exemplary embodiment of this disclosure.

[0036] Figure 8 This diagram illustrates a comparison of fairness indices under different entanglement swap success rates in an exemplary embodiment of this disclosure.

[0037] Figure 9 This diagram illustrates a comparison of fairness indices for different numbers of SD pairs in an exemplary embodiment of this disclosure.

[0038] Figure 10 This diagram illustrates a comparison of request service rates under different fidelity thresholds in an exemplary embodiment of this disclosure.

[0039] Figure 11 This diagram illustrates a comparison of request service rates under different entanglement swap success rates in an exemplary embodiment of this disclosure.

[0040] Figure 12 This diagram illustrates a comparison of request service rates for different numbers of SD pairs in an exemplary embodiment of this disclosure. Detailed Implementation

[0041] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that this disclosure will be more comprehensive and complete, and will fully convey the concept of the exemplary embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0042] Furthermore, the accompanying drawings are merely illustrative of this disclosure and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0043] This example implementation first provides a method for entanglement request scheduling based on quantum networks, referencing... Figure 1-2 As shown, the method includes:

[0044] S1: Based on the entangled connection established between adjacent quantum nodes, a routing path selection method for a single long-distance request is established to select the routing path with the highest end-to-end fidelity for each request.

[0045] S2: Establish a two-stage request scheduling strategy, which includes a resource pre-allocation stage and a resource replenishment stage.

[0046] S21: During the resource pre-allocation phase, requests in the current request round are executed in descending order of end-to-end fidelity.

[0047] S22: For requests whose requirements are not met in the resource pre-allocation phase, enter the resource supplementation phase, and try to execute the suboptimal route path in the resource supplementation phase.

[0048] S3: Establish a comprehensive evaluation mechanism based on network throughput, fairness index, and request service rate.

[0049] In this embodiment, the technical solution uses end-to-end fidelity as a key indicator for request priority ranking in the quantum network purification operation scenario. By establishing a refined fidelity evaluation model, the resource requirements and application value of different entanglement requests are accurately quantified. High-fidelity requests are prioritized for entry into the resource allocation queue. At the same time, the resource loss characteristics during the purification process are dynamically adapted to achieve optimal allocation of entanglement resources in both time and space dimensions. This ensures the service quality of high-value requests while avoiding resource idleness and waste, significantly improving the overall resource utilization rate.

[0050] The method described above in this example implementation will now be explained in more detail.

[0051] In one embodiment, in S1, based on the entanglement established between neighboring quantum nodes, a single long-distance request establishes end-to-end entanglement. First, a routing path needs to be determined. In a quantum network, for a set of requests arriving within a single time slot, different requests require different amounts of entanglement resources.

[0052] To address the competition among these requests, and considering the limited entanglement resources that cannot satisfy all requests, each request needs an optimal routing path to minimize entanglement resource consumption, including purification operations. In this embodiment, the routing path with the highest end-to-end fidelity is selected for each request, thereby freeing up as many entanglement resources as possible. This not only improves the accuracy of resource requirements in relation to application value but also reduces resource consumption.

[0053] Furthermore, in S2, in order to improve the utilization of entangled resources and serve more requests, this embodiment confirms a two-stage request scheduling strategy that includes a resource pre-allocation stage and a resource replenishment stage, in order to determine the execution order of different requests after the routing path of each request is determined.

[0054] Furthermore, during the resource pre-allocation phase, the end-to-end fidelity of all requests within the current request round is determined.

[0055] The request with the highest end-to-end fidelity is executed first, and the remaining requests are scheduled and executed in descending order of end-to-end fidelity.

[0056] This phase prioritizes requests with high fidelity, which can be processed without requiring probabilistic quantum operations that consume inherent resources, such as entanglement purification, thus achieving optimal resource utilization.

[0057] Furthermore, during the resource replenishment phase, requests whose needs were not met during the resource pre-allocation phase are combined with newly arriving request rounds to form a new request set, and the unmet requests are executed first.

[0058] This operation avoids excessively long response times, thus ensuring fairness.

[0059] Furthermore, to avoid excessively long response times and ensure fairness, the resource replenishment phase executes each unmet request only once. If the previous round of requests still cannot meet the demand, it will not participate in subsequent scheduling processes.

[0060] When there is more than one unmet request, the execution order of these requests in the resource replenishment phase is consistent with that in the resource pre-allocation phase.

[0061] Furthermore, to further save time and ensure fairness, requests in the newly arrived request round are scheduled and executed sequentially in descending order of end-to-end fidelity.

[0062] When the final request round of the resource replenishment phase is completed, if new unmet requests arise, the resource replenishment phase will be restarted for these requests.

[0063] Furthermore, let's illustrate S2 with an example:

[0064] like Figure 3 As shown, in this embodiment, it is assumed that two rounds of entanglement request sets need to be processed in each time slot, and each request set contains 3 entanglement requests, namely three source-destination pairs (SD pairs).

[0065] During the resource pre-allocation phase, the SD pair (s2, d2) with the highest end-to-end fidelity in the first round of requests is first identified. Then, all requests are sorted by fidelity and executed sequentially. This phase prioritizes requests with higher fidelity; however, due to limited resources, the needs of some SD pairs, such as SD pair (s1, d1), cannot be met. These requests will be filtered out and moved to the next round of requests, entering the resource replenishment phase. SD pairs whose needs are already met during the resource pre-allocation phase will not participate in the resource replenishment phase.

[0066] During the resource replenishment phase, the unmet SD pair (s1, d1) will form a new request set with the newly arrived request rounds. In this embodiment, these failed requests will be executed first, regardless of whether the SD pair (s1, d1) has the highest end-to-end fidelity. The remaining requests will still be sorted by fidelity and wait to be executed in sequence. For the SD pair (s1, d1), since the optimal route path selected in the first phase has failed, this embodiment will select the second-highest fidelity route path for a second attempt. If the request SD pair (s1, d1) still cannot meet the requirements, it will no longer participate in the subsequent scheduling process. If the previous request set R1 generated multiple unmet requests, the execution order of these priority-executed requests in the next request set R2 will remain consistent with the previous round.

[0067] Furthermore, in S3, within the comprehensive evaluation mechanism, for each individual request, throughput is defined as the number of end-to-end entangled connections established for that request.

[0068] When the selected route path does not require purification, the throughput is determined by the link capacity of this route path.

[0069] In this step, it is assumed that the routing path contains If the success rate of hop and entanglement swap is h, then the path throughput is... It can be calculated as:

[0070] ,

[0071] in Represents the i-th node v i With the j-th node v j Link capacity between them.

[0072] Conversely, when the selected route path requires purification, based on a purification model involving multiple attempts, the throughput achieved by the request on the selected path is [data missing]. It can be calculated as:

[0073] ,

[0074] In the formula, M represents the number of purification attempts. The probability of success for each attempt. This indicates the number of entangled pairs consumed in the purification operation. =2M. Based on the above considerations, the overall network throughput... , where |R| is the number of concurrent source-destination pairs, that is, the total number of requests entangled in a set of requests.

[0075] Furthermore, to evaluate the fairness of different resource scheduling schemes, Jain's index, also known as the fairness index, is used to measure their performance. Fairness Index It can be represented as:

[0076] ,

[0077] Finally, to further evaluate the satisfaction of the entangled request set, the request service rate was... Defined as:

[0078] ,

[0079] In the formula, Representative request r Actual resource requirements.

[0080] To assess the general applicability of this application, the proposed request scheduling algorithm was evaluated using a Waxman stochastic topology model selected by Python simulation software.

[0081] To simulate the impact of environmental noise, a phase-damped noise channel is considered, and the initial phase-damping coefficient p of the quantum link follows a p~N(0.07, 0.01). Within each time slot, the total demand for each round of request sets is 100, randomly generated from each SD pair. As control variables, the fidelity threshold is set to 0.80 by default, the entanglement swap success rate to 0.90, and the number of concurrent SD pairs to 5. Furthermore, the simulation assumes that two rounds of entanglement request sets need to be processed in each time slot, the network has 40 nodes, a link capacity of 50, and 3 attempts at entanglement purification. Finally, the average results after running for 100 time slots are presented.

[0082] Regarding the selection of comparison schemes, two scheduling strategies were chosen: First-In-First-Out (FIFO) and Expected Throughput First (ETF). The FIFO strategy schedules requests based on their arrival time, prioritizing earlier-arriving requests. ETF operates in a greedy mode, always prioritizing requests with the highest expected throughput. Despite lacking resource replenishment capabilities, FIFO and ETF represent two typical scheduling methods, respectively embodying the pursuit of fairness and throughput in scheduling algorithms. They are crucial benchmark algorithms for solving quantum network scheduling problems.

[0083] First, this application analyzes the network throughput performance of three scheduling schemes under different scenarios. The horizontal axis represents line graphs of different parameters for different scheduling algorithms, namely link fidelity, entanglement switching success rate, and number of SD pairs, while the vertical axis represents throughput.

[0084] Figure 4 The impact of different link fidelity thresholds is demonstrated. It can be observed that the network throughput of all algorithms decreases as the fidelity threshold increases. This is because at high fidelity thresholds, the algorithms perform purification operations more frequently, and the additional resource consumption reduces the number of entangled connections that can be established. Compared to the FIFO and ETF algorithms, the proposed algorithm exhibits superior performance due to its two-stage scheduling strategy. The pre-allocation stage prioritizes requests with higher fidelity, minimizing the consumption of entanglement resources by purification operations, thus serving more requests; the resource replenishment stage provides failed requests with additional opportunities to try suboptimal paths, also improving network throughput. The ETF algorithm prioritizes requests with higher expected throughput, an inherent characteristic that allows it to outperform the FIFO algorithm in throughput performance. Furthermore, the differences among the three algorithms become more significant when the fidelity threshold decreases. This is because when the fidelity threshold increases, purification operations consume more resources, leading to a reduction in available network scheduling resources and weakening the algorithm's scheduling capability. Figure 5The results show a positive correlation between entanglement swap success rate and throughput performance; a higher entanglement swap success rate allows for the establishment of more end-to-end entanglements, thereby improving network throughput. Furthermore, the proposed algorithm exhibits a more significant performance advantage when the entanglement swap success rate is high, as the available entanglement resources in the network are more abundant at this point. Figure 6 The impact of the number of SD pairs on network throughput was analyzed. When the number of SD pairs is small, network resource utilization is insufficient, resulting in low throughput for all three algorithms. As the number of SD pairs increases, network resources are fully utilized, and throughput shows an upward trend. However, due to limited network resources, the rate of increase gradually slows down. When the number of SD pairs is large, this application, with its scheduling strategy, can more effectively handle scheduling problems in high-concurrency scenarios, showing a significant advantage over ETF and FIFO algorithms.

[0085] Secondly, this application evaluates the fairness of different algorithms. For example... Figures 7-9 As shown, the horizontal bar chart represents the different parameters in different scheduling algorithms, namely link fidelity, entanglement switching success rate, and number of SD pairs, while the vertical bar chart represents the fairness index.

[0086] like Figure 7 As shown, Jain's index generally decreases with increasing fidelity threshold. This is because increasing the fidelity threshold reduces the entangled link resources available for each request on the selected path, thus lowering Jain's index. When the fidelity threshold reaches 0.90, purification operations become almost inevitable, making it difficult to satisfy the needs of each request. This "relative fairness" causes a slight increase in Jain's index. Among the three algorithms, the proposed algorithm exhibits the best fairness due to its more rational resource utilization. Since FIFO processes requests sequentially, its fairness is superior to the ETF algorithm, but this advantage gradually diminishes as available resources decrease. Figure 8 The impact of entanglement swap success rate on fairness is demonstrated. A higher entanglement swap success rate results in a larger Jain's exponent, because an increased success rate leads to more end-to-end entangled pairs being established per request, thus increasing the Jain's exponent. Compared to the "greedy" ETF algorithm, both the proposed algorithm and the FIFO algorithm exhibit superior fairness. Figure 9 This paper demonstrates the variation of Jain's exponent with the number of SD pairs. As the number of SD pairs increases, intensified resource competition reduces the resources available for each request, causing Jain's exponent to decrease. However, when the number of SD pairs is high, Jain's exponent actually increases slightly. This is because in high-concurrency scenarios, since this application assumes a fixed total network demand, the resource requirement of each SD pair will decrease, thus achieving "relative fairness." Therefore, when the number of SD pairs is low, the excessively high resource requirement of each SD pair will hinder the full realization of the fairness advantage of the FIFO algorithm.

[0087] Finally, this application explores the request service rates of three scheduling algorithms in different scenarios. Figures 10-12 As shown, the horizontal bar chart represents the different parameters in different scheduling algorithms, namely link fidelity, entanglement switching success rate, and number of SD pairs, while the vertical bar chart represents the request service rate.

[0088] like Figure 10 As shown, as the link fidelity threshold increases, the purification operation reduces available network resources, leading to a gradual decrease in the request service rate. This application, with its two-stage request scheduling strategy, can handle more requests. For the ETF and FIFO algorithms, their ability to handle parallel requests is essentially equivalent. Figure 11 As shown, the higher success rate of entanglement swaps further improves the performance of each scheduling algorithm, which is consistent with the previous analysis in this application. Compared with the ETF and FIFO algorithms, this algorithm shows greater stability. Figure 12 This demonstrates the impact of different numbers of SD pairs on the request service rate. When the number of SD pairs is small, a single request requires a large amount of resources (total network demand is fixed), making it difficult for the selected path to meet the demand, resulting in a low request service rate. As the number of SD pairs increases, network resources are fully utilized, and the request service rate gradually rises and tends to saturate.

[0089] These results demonstrate that, compared to traditional ETF and FIFO algorithms, this application has significant advantages in quantum network request scheduling. The proposed algorithm can provide higher network throughput and request service rate while ensuring a certain degree of fairness.

[0090] It should be noted that although the various steps for the algorithm execution flow are mentioned in the detailed description above, this division is not mandatory. In fact, according to the embodiments of this application, the features and functions of two or more steps described above can be embodied in a single step. Conversely, the features and functions of one step described above can be further divided into multiple steps for embodiment. Some or all of the steps can be selected to achieve the purpose of this application according to actual needs. Those skilled in the art can understand and implement this without any inventive effort.

[0091] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein.

Claims

1. A method for entanglement request scheduling based on quantum networks, characterized in that, Includes the following steps: Based on the establishment of entangled connections between adjacent quantum nodes, a routing path selection method for a single long-distance request is established to select the routing path with the highest end-to-end fidelity for each request. A two-stage request scheduling strategy is established, which includes a resource pre-allocation stage and a resource replenishment stage. During the resource pre-allocation phase, requests within the current request round are executed in descending order of end-to-end fidelity. For requests whose requirements are not met in the resource pre-allocation phase, the process proceeds to the resource replenishment phase, where a suboptimal route is selected for execution. Establish a comprehensive evaluation mechanism based on network throughput, fairness index, and request service rate; In the comprehensive evaluation mechanism, for each individual request, the throughput is defined as the number of end-to-end entangled connections established by that request, wherein the throughput is determined by the link capacity of the selected route path when the selected route path does not require purification operations. In the comprehensive evaluation mechanism, the fairness index adopts Jain's index; In the aforementioned comprehensive evaluation mechanism, the request service rate is defined as: ， in, For the requested service rate, Let r be an element in R, representing the throughput achieved by this request on the selected path. This represents the actual resource requirements of request r. This refers to the number of concurrent source-destination pairs, which is the total number of requests entangled in a set of requests.

2. The method for entanglement-based request scheduling in quantum networks according to claim 1, characterized in that, During the resource pre-allocation phase, the end-to-end fidelity of all requests within the current request round is determined; The request with the highest end-to-end fidelity is executed first, and the remaining requests are scheduled and executed in descending order of end-to-end fidelity.

3. The method for entanglement-based request scheduling in quantum networks according to claim 1, characterized in that, During the resource replenishment phase, requests whose needs were not met during the resource pre-allocation phase are combined with newly arriving request rounds to form a new request set, and the unmet requests are executed first.

4. The method for entanglement-based request scheduling in quantum networks according to claim 3, characterized in that, The resource replenishment phase is executed only once for each of the unmet needs.

5. The method for entanglement request scheduling based on quantum networks according to claim 3, characterized in that, When there is more than one unmet request, the execution order of the multiple unmet requests in the resource replenishment phase is consistent with that in the resource pre-allocation phase.

6. The method for entanglement-based request scheduling in quantum networks according to claim 3, characterized in that, The requests in the newly arriving request round are scheduled and executed sequentially in descending order of end-to-end fidelity.

7. The method for entanglement-based request scheduling in quantum networks according to claim 3, characterized in that, When the final request round of the resource replenishment phase is completed, if a new unmet request arises, the resource replenishment phase will be restarted for the new unmet request.

Images