CN115843070B - Ocean sensing network calculation unloading method and system based on energy collection technology - Google Patents

Ocean sensing network calculation unloading method and system based on energy collection technology Download PDF

Info

Publication number
CN115843070B
CN115843070B CN202310152246.XA CN202310152246A CN115843070B CN 115843070 B CN115843070 B CN 115843070B CN 202310152246 A CN202310152246 A CN 202310152246A CN 115843070 B CN115843070 B CN 115843070B
Authority
CN
China
Prior art keywords
time slot
task
calculation
marine
model
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202310152246.XA
Other languages
Chinese (zh)
Other versions
CN115843070A (en
Inventor
张玮
林迅宇
史慧玲
郝昊
丁伟
谭立状
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Qilu University of Technology
National Supercomputing Center in Jinan
Original Assignee
Qilu University of Technology
National Supercomputing Center in Jinan
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Qilu University of Technology, National Supercomputing Center in Jinan filed Critical Qilu University of Technology
Priority to CN202310152246.XA priority Critical patent/CN115843070B/en
Publication of CN115843070A publication Critical patent/CN115843070A/en
Application granted granted Critical
Publication of CN115843070B publication Critical patent/CN115843070B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D30/00—Reducing energy consumption in communication networks
    • Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks

Landscapes

  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Testing Or Calibration Of Command Recording Devices (AREA)

Abstract

The invention discloses a calculation unloading method and a calculation unloading system for a marine sensing network based on an energy collection technology, which relate to the technical field of marine observation sensing networks.

Description

基于能量收集技术的海洋传感网络计算卸载方法及系统Ocean sensor network computing offloading method and system based on energy harvesting technology

技术领域Technical Field

本发明涉及海洋观测传感网络技术领域，尤其涉及一种基于能量收集技术的海洋传感网络计算卸载方法及系统。The present invention relates to the technical field of ocean observation sensor networks, and in particular to an ocean sensor network computing offloading method and system based on energy harvesting technology.

背景技术Background Art

传统的海洋观测主要是以调查船、潜浮标为主的海基观测或以卫星遥感、航空观测为基础的天基观测。由于海洋环境的复杂性和独特性，海洋观测数据存在的短暂、不连续等问题一直制约着海洋科学的发展。源自冷战时期美国海军水声监视系统的海底观测网是人类建立的第三种海洋科学观测平台。在现代传感器、水下机器人、海底光纤电缆、物联网、大数据等新型技术的推动下，海底观测网融合物理海洋、海洋化学、海洋地球物理、海洋生态等学科，解决深海、极端环境下高分辨率和实时获取海洋观测数据的技术难题，可以深入到海洋内部观测和认识海洋，实现从海底到海面全天候、长期、连续、综合、实时、原位观测。Traditional ocean observations are mainly sea-based observations based on survey ships and submerged buoys or space-based observations based on satellite remote sensing and aerial observations. Due to the complexity and uniqueness of the marine environment, the short-term and discontinuous nature of marine observation data has always restricted the development of marine science. The seabed observation network, which originated from the US Navy's hydroacoustic surveillance system during the Cold War, is the third marine science observation platform established by humans. Driven by new technologies such as modern sensors, underwater robots, submarine fiber optic cables, the Internet of Things, and big data, the seabed observation network integrates disciplines such as physical oceanography, marine chemistry, marine geophysics, and marine ecology to solve the technical problems of high-resolution and real-time acquisition of marine observation data in deep sea and extreme environments. It can go deep into the ocean to observe and understand the ocean, and realize all-weather, long-term, continuous, comprehensive, real-time, and in-situ observations from the seabed to the sea surface.

建立分布式、网络化、互动式、综合性智能立体观测网是海洋科学观测的发展趋势。随着物联网技术在海洋领域的应用，通过统一、通用的数据标准整合分散在各处的观测站、观测节点、卫星遥感、无人水面艇等观测手段进行协同工作，形成覆盖近岸、区域及全球海域的层次化、综合化与智能化的空-天-海洋一体化立体观测网络。由于传感设备的资源（如计算和能源容量）有限，当能源较低时，计算能力会大幅下降，甚至会导致业务失败。通过传感网网关将传感设备中的计算任务卸载到远程云，可以有效解决计算问题。然而，云处理无法满足诸多观测应用的严格延迟要求，这些应用通常需要实时处理和响应（例如灾难响应）。Establishing a distributed, networked, interactive, comprehensive and intelligent three-dimensional observation network is the development trend of marine scientific observation. With the application of Internet of Things technology in the marine field, observation stations, observation nodes, satellite remote sensing, unmanned surface vessels and other observation methods scattered in various places are integrated through unified and universal data standards to work together to form a hierarchical, integrated and intelligent air-space-ocean integrated three-dimensional observation network covering nearshore, regional and global waters. Due to the limited resources (such as computing and energy capacity) of sensor equipment, when the energy is low, the computing power will drop significantly, and even cause business failure. Offloading the computing tasks in the sensor equipment to the remote cloud through the sensor network gateway can effectively solve the computing problem. However, cloud processing cannot meet the strict latency requirements of many observation applications, which usually require real-time processing and response (such as disaster response).

雾节点是雾体系结构的基本元素，雾节点可以是任何提供雾架构的计算、网络、存储和加速元素的设备，例如，工业控制器、交换机、路由器、嵌入式服务器、复杂网关、可编程逻辑控制器（PLC，ProgrammableLogic Controller）以及智能物联网节点（如视频监控摄像机）等。考虑到雾节点具有计算、存储网络资源的功能，因此，通过在海洋传感设备附近部署雾节点，可以有效提高海洋观测网的服务性能。Fog nodes are the basic elements of fog architecture. Fog nodes can be any device that provides computing, networking, storage, and acceleration elements of fog architecture, such as industrial controllers, switches, routers, embedded servers, complex gateways, programmable logic controllers (PLCs), and smart IoT nodes (such as video surveillance cameras). Considering that fog nodes have the functions of computing and storing network resources, the service performance of the ocean observation network can be effectively improved by deploying fog nodes near marine sensing equipment.

雾节点通常连接在传感网网关上，用于处理来自传感网设备的计算任务，以提供即时的业务响应；能量收集技术可以收集环境中容易获得的少量非传统能量并将之转化为电能，为传感设备持续供电。然而，在环境复杂、拓扑动态变化的海洋环境中，传统任务分配方式无法有效应用，由于服务质量（QoS，Quality of Service）的要求，传统卸载方法要么将所有计算任务卸载到最近的雾节点，这会导致一些雾节过载而传感设备负载不足；要么由传感设备承担计算任务，使得传感设备电量快速下降导致任务失败。因此，针对节点动态变化的海洋环境，现有的任务分配/卸载方法无法适用于应用了能量收集技术的海洋传感网络。Fog nodes are usually connected to sensor network gateways to process computing tasks from sensor network devices to provide immediate business responses; energy harvesting technology can collect small amounts of non-traditional energy that is easily available in the environment and convert it into electrical energy to continuously power sensor devices. However, in marine environments with complex environments and dynamically changing topologies, traditional task allocation methods cannot be effectively applied. Due to the requirements of Quality of Service (QoS), traditional offloading methods either offload all computing tasks to the nearest fog node, which will cause some fog nodes to be overloaded and sensor devices to be underloaded; or the sensor devices will take on the computing tasks, causing the battery power of the sensor devices to drop rapidly, resulting in task failure. Therefore, for marine environments with dynamically changing nodes, existing task allocation/offloading methods are not applicable to marine sensor networks that use energy harvesting technology.

发明内容Summary of the invention

为解决上述现有技术的不足，本发明提供了一种基于能量收集技术的海洋传感网络计算卸载方法及系统，适用于应用了能量收集技术的海洋传感网络，提供最优计算卸载策略，实现为海洋传感设备和雾节点确定合适的任务卸载比例，提高海洋传感网络的计算性能，避免节点资源浪费。In order to solve the deficiencies of the above-mentioned prior art, the present invention provides a method and system for offloading computing of an ocean sensor network based on energy harvesting technology, which is suitable for an ocean sensor network that uses energy harvesting technology, provides an optimal computing offloading strategy, and determines a suitable task offloading ratio for ocean sensor equipment and fog nodes, thereby improving the computing performance of the ocean sensor network and avoiding waste of node resources.

第一方面，本公开提供了一种基于能量收集技术的海洋传感网络计算卸载方法。In a first aspect, the present disclosure provides a method for offloading computation in an ocean sensor network based on energy harvesting technology.

一种基于能量收集技术的海洋传感网络计算卸载方法，包括：A method for offloading computation of an ocean sensor network based on energy harvesting technology, comprising:

在每一时隙内，获取当前时隙海洋传感设备的实时信息；所述实时信息包括海洋传感设备所请求计算任务的任务请求指示值、海洋传感设备访问雾节点的信道功率以及海洋传感设备的电池能量水平；In each time slot, real-time information of the ocean sensor device in the current time slot is obtained; the real-time information includes a task request indication value of the computing task requested by the ocean sensor device, a channel power for the ocean sensor device to access the fog node, and a battery energy level of the ocean sensor device;

构建海洋传感设备在当前时隙内执行计算任务的计算延迟和能量消耗，以及将计算任务卸载至雾节点的传输时延和能量消耗；Construct the computational delay and energy consumption of ocean sensor devices performing computational tasks in the current time slot, as well as the transmission delay and energy consumption of offloading computational tasks to fog nodes;

搭建计算任务执行成本模型、海洋传感设备能耗模型和能量收集模型，以长期平均执行成本最小为目标函数，结合所搭建的模型，构建计算卸载模型；Build a computing task execution cost model, a marine sensor equipment energy consumption model, and an energy collection model. Take the minimum long-term average execution cost as the objective function, and build a computing offloading model based on the built models.

根据所获取的实时信息，利用动态算法求解计算卸载模型，求解获得当前时隙的最优计算卸载决策。According to the real-time information obtained, the dynamic algorithm is used to solve the computation offloading model to obtain the optimal computation offloading decision for the current time slot.

进一步的技术方案，基于在第t个时隙时海洋传感设备完成计算任务所需的W个CPU周期的预定频率

Figure SMS_1
，构建海洋传感设备在第t个时隙内本地执行计算任务的计算延迟
Figure SMS_2
和能量消耗
Figure SMS_3
，公式为：A further technical solution is based on the predetermined frequency of W CPU cycles required for the ocean sensing device to complete the computing task at the tth time slot.
Figure SMS_1
, construct the computational delay of the ocean sensor device to perform the computational task locally in the tth time slot
Figure SMS_2
and energy consumption
Figure SMS_3
, the formula is:

Figure SMS_4
Figure SMS_4

Figure SMS_5
Figure SMS_5

其中，k为有效开关电容，w表示CPU周期个数，w=1,…,W。Where, k is the effective switching capacitance, w represents the number of CPU cycles, w=1,…,W.

进一步的技术方案，基于海洋传感设备访问雾节点的信道功率

Figure SMS_6
和发射功率
Figure SMS_7
，根据香农公式，构建海洋传感设备在第t个时隙内将计算任务卸载至雾节点的传输时延
Figure SMS_8
和能量消耗
Figure SMS_9
，公式为：A further technical solution is to access the channel power of fog nodes based on ocean sensing equipment
Figure SMS_6
and transmit power
Figure SMS_7
According to Shannon’s formula, the transmission delay of the ocean sensor device to offload the computing task to the fog node in the tth time slot is constructed.
Figure SMS_8
and energy consumption
Figure SMS_9
, the formula is:

Figure SMS_10
Figure SMS_10

Figure SMS_11
Figure SMS_11

其中，L表示计算任务的输入数据量，r表示第t个时隙的传输速率，且Where L represents the input data volume of the computing task, r represents the transmission rate of the tth time slot, and

Figure SMS_12
Figure SMS_12

上式中，ω表示信道带宽，σ表示信道内部的高斯噪声功率。In the above formula, ω represents the channel bandwidth, and σ represents the Gaussian noise power inside the channel.

进一步的技术方案，将执行延迟和任务丢弃成本的加权和作为计算任务执行成本，则计算任务执行成本模型为：A further technical solution is to use the weighted sum of execution delay and task abandonment cost as the calculation task execution cost, and then the calculation task execution cost model is:

Figure SMS_13
Figure SMS_13

其中，

Figure SMS_14
表示第t个时隙的计算任务执行成本，
Figure SMS_15
表示获取到计算任务后却未执行的惩罚，
Figure SMS_16
表示海洋传感设备在第t个时隙内本地执行计算任务的计算延迟，
Figure SMS_17
表示海洋传感设备在第t个时隙内将计算任务卸载至雾节点的传输时延，
Figure SMS_18
=1和
Figure SMS_19
=1分别表示在第t个时隙时，请求的计算任务在海洋传感设备上执行和计算任务卸载到雾节点。in,
Figure SMS_14
represents the execution cost of the computational task in the tth time slot,
Figure SMS_15
It indicates the penalty for not executing the computing task after obtaining it.
Figure SMS_16
represents the computational delay of the ocean sensor device in executing the computational task locally in the tth time slot,
Figure SMS_17
represents the transmission delay of the ocean sensor device offloading the computing task to the fog node in the tth time slot,
Figure SMS_18
=1 and
Figure SMS_19
=1 respectively indicates that at the tth time slot, the requested computing task is executed on the ocean sensor device and the computing task is offloaded to the fog node.

进一步的技术方案，在每个时隙中，将基于能量收集技术到达海洋传感设备的部分能量记为

Figure SMS_20
，构建能量收集模型，为：In a further technical solution, in each time slot, the part of the energy reaching the ocean sensing device based on the energy harvesting technology is recorded as
Figure SMS_20
, construct the energy harvesting model as:

Figure SMS_21
Figure SMS_21

其中，

Figure SMS_22
表示第t个时隙开始时到达海洋传感设备的能量。in,
Figure SMS_22
represents the energy reaching the ocean sensing device at the beginning of the tth time slot.

进一步的技术方案，所述海洋传感设备能耗模型为海洋传感设备在第t个时隙内所消耗的能量，公式为：In a further technical solution, the energy consumption model of the ocean sensing device is the energy consumed by the ocean sensing device in the tth time slot, and the formula is:

Figure SMS_23
Figure SMS_23

且满足能量关系约束：And satisfy the energy relationship constraints:

Figure SMS_24
Figure SMS_24

其中，

Figure SMS_27
表示第t个时隙时海洋传感设备的电池能量水平，
Figure SMS_28
表示计算卸载模型的指标，
Figure SMS_31
=1和
Figure SMS_26
=1分别表示在第t个时隙时，请求的计算任务在海洋传感设备上执行和计算任务卸载到雾节点，
Figure SMS_29
表示当前第t个时隙时CPU周期的预定频率，
Figure SMS_30
表示海洋传感设备访问雾节点的发射功率，
Figure SMS_32
表示海洋传感设备在第t个时隙内本地执行计算任务的能量消耗，
Figure SMS_25
表示海洋传感设备在第t个时隙内将计算任务卸载至雾节点的能量消耗。in,
Figure SMS_27
represents the battery energy level of the ocean sensing device at the tth time slot,
Figure SMS_28
Indicates the index for calculating the offloading model,
Figure SMS_31
=1 and
Figure SMS_26
=1 respectively indicates that at the tth time slot, the requested computing task is executed on the ocean sensor device and the computing task is offloaded to the fog node,
Figure SMS_29
Indicates the expected frequency of the CPU cycle at the current t-th time slot,
Figure SMS_30
represents the transmission power of the ocean sensor device accessing the fog node,
Figure SMS_32
represents the energy consumption of the ocean sensor device performing the computing task locally in the tth time slot,
Figure SMS_25
It represents the energy consumption of the ocean sensor device when offloading the computing task to the fog node in the tth time slot.

进一步的技术方案，所述计算卸载模型为：In a further technical solution, the calculation offloading model is:

Figure SMS_33
Figure SMS_33

s.t.s.t.

Figure SMS_34
Figure SMS_34

Figure SMS_35
Figure SMS_35

Figure SMS_36
Figure SMS_36

Figure SMS_37
Figure SMS_37

Figure SMS_38
Figure SMS_38

Figure SMS_39
Figure SMS_39

其中，

Figure SMS_42
表示第t个时隙时的最大传输功率，
Figure SMS_43
表示第t个时隙时最大CPU周期的预定频率，
Figure SMS_48
表示第t个时隙的计算任务执行成本，t=0,1,...,T-1，T表示时隙的总个数，
Figure SMS_41
表示当前第t个时隙时海洋传感设备的电池能量水平，
Figure SMS_44
表示计算卸载模型的指标，
Figure SMS_49
=1和
Figure SMS_50
=1分别表示在第t个时隙时，请求的计算任务在海洋传感设备上执行和计算任务卸载到雾节点，
Figure SMS_40
表示当前第t个时隙时CPU周期的预定频率，
Figure SMS_45
表示海洋传感设备访问雾节点的发射功率，
Figure SMS_46
表示到达海洋传感设备的部分能量，
Figure SMS_47
表示第t个时隙时海洋传感设备完成计算任务所需的W个CPU周期的预定频率。in,
Figure SMS_42
represents the maximum transmission power at the tth time slot,
Figure SMS_43
represents the predetermined frequency of the maximum CPU cycle at the tth time slot,
Figure SMS_48
represents the execution cost of the computational task in the tth time slot, t=0,1,...,T-1, T represents the total number of time slots,
Figure SMS_41
represents the battery energy level of the ocean sensor device at the current t-th time slot,
Figure SMS_44
Indicates the index for calculating the offloading model,
Figure SMS_49
=1 and
Figure SMS_50
=1 respectively indicates that at the tth time slot, the requested computing task is executed on the ocean sensor device and the computing task is offloaded to the fog node,
Figure SMS_40
Indicates the expected frequency of the CPU cycle at the current t-th time slot,
Figure SMS_45
represents the transmission power of the ocean sensor device accessing the fog node,
Figure SMS_46
represents the portion of energy reaching the ocean sensing device,
Figure SMS_47
Represents the predetermined frequency of W CPU cycles required for the ocean sensor device to complete the computing task at the tth time slot.

进一步的技术方案，利用动态算法求解当前时隙内的最优计算卸载决策，包括：A further technical solution uses a dynamic algorithm to solve the optimal computing offloading decision in the current time slot, including:

构建李雅普诺夫函数；Construct Lyapunov functions;

根据李雅普诺夫函数，构建李雅普诺夫漂移函数和李雅普诺夫漂移加惩罚函数；According to the Lyapunov function, construct the Lyapunov drift function and the Lyapunov drift plus penalty function;

根据李雅普诺夫漂移函数和李雅普诺夫漂移加惩罚函数，化简计算卸载模型；According to the Lyapunov drift function and the Lyapunov drift plus penalty function, the unloading model is simplified and calculated;

对化简后的计算卸载模型进行求解，得到当前时隙内的最优计算卸载决策。The simplified computation offloading model is solved to obtain the optimal computation offloading decision in the current time slot.

进一步的技术方案，所述求解包括：A further technical solution is that the solution includes:

在第t个时隙开始时，获取当前时隙海洋传感设备所请求计算任务的任务请求指示值

Figure SMS_51
、海洋传感设备的电池能量水平
Figure SMS_52
和海洋传感设备访问雾节点的信道功率
Figure SMS_53
；At the beginning of the tth time slot, obtain the task request indication value of the computing task requested by the ocean sensing device in the current time slot
Figure SMS_51
, Battery energy levels of ocean sensing devices
Figure SMS_52
and channel power for ocean sensor devices to access fog nodes
Figure SMS_53
;

根据所获取的当前时隙的信息，代入化简后的计算卸载模型中进行求解，确定当前时隙的计算卸载模型的指标

Figure SMS_54
、CPU周期的预定频率
Figure SMS_55
、海洋传感设备访问雾节点的发射功率
Figure SMS_56
基于能量收集技术到达海洋传感设备的部分能量
Figure SMS_57
；According to the information of the current time slot obtained, substitute it into the simplified calculation offloading model for solution to determine the index of the calculation offloading model of the current time slot
Figure SMS_54
, the predetermined frequency of CPU cycles
Figure SMS_55
, Transmission power of marine sensor equipment accessing fog nodes
Figure SMS_56
Part of the energy reaching ocean sensing devices based on energy harvesting technology
Figure SMS_57
;

根据计算确定的结果更新海洋传感设备的电池能量队列，并判断是否达到当前时隙的稳定值，若稳定，则将计算确定的结果作为当前时隙内的最优计算卸载决策输出，否则计算下一时隙的最优计算卸载决策。The battery energy queue of the ocean sensor equipment is updated according to the calculated results, and it is determined whether the stable value of the current time slot is reached. If stable, the calculated results are output as the optimal computing unloading decision in the current time slot, otherwise the optimal computing unloading decision for the next time slot is calculated.

第二方面，本公开提供了一种基于能量收集技术的海洋传感网络计算卸载系统。In a second aspect, the present disclosure provides a marine sensor network computing offloading system based on energy harvesting technology.

一种基于能量收集技术的海洋传感网络计算卸载系统，包括：A marine sensor network computing unloading system based on energy harvesting technology, comprising:

信息获取模块，用于在每一时隙内，获取当前时隙海洋传感设备的实时信息，所述实时信息包括海洋传感设备所请求计算任务的任务请求指示值、海洋传感设备访问雾节点的信道功率以及海洋传感设备的电池能量水平；An information acquisition module is used to acquire real-time information of the ocean sensor device in the current time slot in each time slot, wherein the real-time information includes a task request indication value of the computing task requested by the ocean sensor device, a channel power for the ocean sensor device to access the fog node, and a battery energy level of the ocean sensor device;

模型构建模块，用于构建海洋传感设备在当前时隙内执行计算任务的计算延迟和能量消耗，以及将计算任务卸载至雾节点的传输时延和能量消耗，搭建计算任务执行成本模型、海洋传感设备能耗模型和能量收集模型，以长期平均执行成本最小为目标函数，结合所搭建的模型，构建计算卸载模型；The model building module is used to construct the computational delay and energy consumption of the ocean sensor equipment in executing the computational task in the current time slot, as well as the transmission delay and energy consumption of offloading the computational task to the fog node, build the computational task execution cost model, the ocean sensor equipment energy consumption model and the energy collection model, and build the computational offloading model with the minimum long-term average execution cost as the objective function in combination with the built model;

卸载决策求解模块，用于根据所获取的实时信息，利用动态算法求解计算卸载模型，求解获得当前时隙的最优计算卸载决策。The offloading decision solving module is used to solve the computing offloading model using a dynamic algorithm based on the acquired real-time information, and to obtain the optimal computing offloading decision for the current time slot.

以上一个或多个技术方案存在以下有益效果：One or more of the above technical solutions have the following beneficial effects:

1、本发明提供了一种基于能量收集技术的海洋传感网络计算卸载方法及系统，通过获取海洋传感设备在每一时隙内的实时信息，构建海洋传感设备在该时隙内执行计算任务的计算延迟和能量消耗以及将计算任务卸载至雾节点的传输时延和能量消耗，以此建立计算任务执行成本模型、海洋传感设备能耗模型和能量收集模型，并综合所构建的模型，以长期平均执行成本最小为目标函数，构建计算卸载模型，利用动态算法求解当前时隙内的最优计算卸载决策，实现了为应用能量收集技术的海洋传感网络提供最优计算卸载策略，为海洋传感设备和雾节点确定合适的任务卸载比例，提高海洋传感网络的计算性能，避免了节点资源浪费。1. The present invention provides a method and system for computing offloading of an ocean sensor network based on energy harvesting technology. By acquiring real-time information of an ocean sensor device in each time slot, the computing delay and energy consumption of the ocean sensor device in executing computing tasks in the time slot, as well as the transmission delay and energy consumption of offloading computing tasks to fog nodes are constructed, thereby establishing a computing task execution cost model, an ocean sensor device energy consumption model, and an energy harvesting model. The constructed models are integrated to construct a computing offloading model with the minimum long-term average execution cost as the objective function. A dynamic algorithm is used to solve the optimal computing offloading decision in the current time slot, thereby providing an optimal computing offloading strategy for an ocean sensor network that uses energy harvesting technology, determining a suitable task offloading ratio for ocean sensor devices and fog nodes, improving the computing performance of the ocean sensor network, and avoiding waste of node resources.

2、本发明考虑了海洋传感设备资源有限，无法满足存储最优策略内存需求的问题，提出了一种动态算法进行求解，利用李雅普诺夫漂移函数和李雅普诺夫漂移加惩罚函数化简计算卸载模型，根据获取的实时信息进行迭代求解，获取当前最优计算卸载决策，以此进行资源分配，优化资源分配结果，减少长期平均执行成本，提高系统的用户服务质量。2. The present invention takes into account the problem that the resources of marine sensor equipment are limited and cannot meet the memory requirements for storing the optimal strategy, and proposes a dynamic algorithm for solving it. It uses the Lyapunov drift function and the Lyapunov drift plus penalty function to simplify the calculation offloading model, and iterates the solution based on the real-time information obtained to obtain the current optimal calculation offloading decision, so as to allocate resources, optimize the resource allocation results, reduce the long-term average execution cost, and improve the user service quality of the system.

附图说明BRIEF DESCRIPTION OF THE DRAWINGS

构成本发明的一部分的说明书附图用来提供对本发明的进一步理解，本发明的示意性实施例及其说明用于解释本发明，并不构成对本发明的不当限定。The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

图1为本发明实施例所述基于能量收集技术的海洋传感网络计算卸载方法的流程图；FIG1 is a flow chart of a method for offloading computation in an ocean sensor network based on energy harvesting technology according to an embodiment of the present invention;

图2为本发明实施例中计算卸载模型的示意图；FIG2 is a schematic diagram of a computational offloading model in an embodiment of the present invention;

图3为本发明实施例中能量收集的示意图。FIG. 3 is a schematic diagram of energy collection in an embodiment of the present invention.

具体实施方式DETAILED DESCRIPTION

应该指出，以下详细说明都是示例性的，旨在对本发明提供进一步的说明。除非另有指明，本文使用的所有技术和科学术语具有与本发明所属技术领域的普通技术人员通常理解的相同含义。It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

需要注意的是，这里所使用的术语仅是为了描述具体实施方式，而非意图限制根据本发明的示例性实施方式。如在这里所使用的，除非上下文另外明确指出，否则单数形式也意图包括复数形式，此外，还应当理解的是，当在本说明书中使用术语“包含”和/或“包括”时，其指明存在特征、步骤、操作、器件、组件和/或它们的组合。It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and/or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and/or combinations thereof.

实施例一Embodiment 1

本实施例提供了一种基于能量收集技术的海洋传感网络计算卸载方法，如图1所示，包括：This embodiment provides a method for offloading computation in an ocean sensor network based on energy harvesting technology, as shown in FIG1 , including:

步骤S1、在每一时隙内，获取当前时隙海洋传感设备的实时信息，所述实时信息包括海洋传感设备所请求计算任务的任务请求指示值、海洋传感设备访问雾节点的信道功率以及海洋传感设备的电池能量水平；Step S1, in each time slot, obtaining the real-time information of the ocean sensor device in the current time slot, wherein the real-time information includes the task request indication value of the computing task requested by the ocean sensor device, the channel power of the ocean sensor device accessing the fog node, and the battery energy level of the ocean sensor device;

步骤S2、构建海洋传感设备在当前时隙内执行计算任务的计算延迟和能量消耗，以及将计算任务卸载至雾节点的传输时延和能量消耗；Step S2: construct the computational delay and energy consumption of the ocean sensor device executing the computational task in the current time slot, as well as the transmission delay and energy consumption of offloading the computational task to the fog node;

步骤S3、搭建计算任务执行成本模型、海洋传感设备能耗模型和能量收集模型，以长期平均执行成本最小为目标函数，结合所搭建的模型，构建计算卸载模型；Step S3: construct a computing task execution cost model, a marine sensor equipment energy consumption model, and an energy collection model, and construct a computing offloading model by taking the minimum long-term average execution cost as the objective function and combining the constructed models;

步骤S4、根据所获取的实时信息，利用动态算法求解计算卸载模型，求解获得当前时隙的最优计算卸载决策，其中，该计算卸载决策用以表征海洋传感设备所请求的计算任务的执行方式。Step S4: according to the acquired real-time information, a dynamic algorithm is used to solve the computational offloading model to obtain the optimal computational offloading decision for the current time slot, wherein the computational offloading decision is used to characterize the execution mode of the computational task requested by the ocean sensing device.

本实施例考虑数个拥有计算能力的海洋传感设备和雾节点组成的计算卸载模型，如图2所示，雾节点位于距离海洋传感设备d米处，并与海洋观测网通信连接，该海洋传感设备（如图2所示的海洋传感设备A、B、C）可以是船舶、海岸基站等，可以通过无线信道对雾节点访问。海洋传感设备关联该节点服务器并运行一个虚拟机，执行代表传感设备的计算任务，且通过将部分计算任务转移到该节点执行，可以显著提升服务质量QoS。This embodiment considers a computing offloading model consisting of several ocean sensor devices with computing capabilities and fog nodes. As shown in Figure 2, the fog node is located d meters away from the ocean sensor device and is connected to the ocean observation network. The ocean sensor device (ocean sensor devices A, B, C shown in Figure 2) can be a ship, a coastal base station, etc., and the fog node can be accessed through a wireless channel. The ocean sensor device is associated with the node server and runs a virtual machine to perform computing tasks on behalf of the sensor device, and by transferring part of the computing tasks to the node for execution, the service quality QoS can be significantly improved.

上述步骤S1中，获取海洋传感设备在当前时隙的实时信息。该实时信息包括海洋传感设备所请求计算任务的任务请求指示值

Figure SMS_58
、海洋传感设备访问雾节点的信道功率
Figure SMS_59
以及海洋传感设备的电池能量水平
Figure SMS_60
。根据所获取的信息，求解所构建的计算卸载模型，求解获得每一时隙内的最优计算卸载决策。In the above step S1, the real-time information of the ocean sensing device in the current time slot is obtained. The real-time information includes the task request indication value of the computing task requested by the ocean sensing device.
Figure SMS_58
, Channel power of ocean sensor devices accessing fog nodes
Figure SMS_59
and battery energy levels of ocean sensing devices
Figure SMS_60
According to the acquired information, the constructed computation offloading model is solved to obtain the optimal computation offloading decision in each time slot.

在本实施例中，将时间切片为多个时隙（或时间片），每个时隙的长度为

Figure SMS_63
，标号为Γ∈{0,1,...}，无线信道在每个时隙内保持静态，但在不同时隙之间变化。用A(L，
Figure SMS_67
)表示一个计算任务，其中，L（单位是位，bit）表示该计算任务的数据量大小，而
Figure SMS_70
表示计算任务完成的截至时间，即计算任务A(L，
Figure SMS_64
)需要在时间
Figure SMS_68
内完成。海洋传感设备上运行应用程序，该程序所请求的计算任务被建模为一个伯努利进程。在每个时隙开始时，计算任务A(L，
Figure SMS_71
)有ρ的概率被请求，相应的，则有(1-ρ)的概率未被请求。若计算任务在第t个时隙内被请求时，令任务请求指示值
Figure SMS_73
= 1，否则
Figure SMS_61
= 0，也即P(
Figure SMS_65
=1)=1-P(
Figure SMS_69
=0)=ρ，t∈Γ，P(
Figure SMS_72
=1)=ρ表示计算任务被请求时的概率为ρ，P(
Figure SMS_62
=0)=1-ρ表示计算任务未被请求时的概率为1-ρ。本实施例中，没有缓冲区可用于计算请求排队，且海洋传感设备上运行的应用程序为关注任务执行时间不大于时隙长度的延迟敏感型应用，即
Figure SMS_66
。In this embodiment, the time is sliced into multiple time slots (or time slices), and the length of each time slot is
Figure SMS_63
, labeled Γ∈{0,1,...}, the wireless channel remains static within each time slot, but changes between different time slots.
Figure SMS_67
) represents a computing task, where L (in bits) represents the data size of the computing task, and
Figure SMS_70
Indicates the deadline for the completion of the computing task, that is, the computing task A(L,
Figure SMS_64
) Need to be in time
Figure SMS_68
The application is run on the ocean sensor device, and the computing task requested by the program is modeled as a Bernoulli process. At the beginning of each time slot, the computing task A(L,
Figure SMS_71
) is requested with probability ρ, and accordingly, it is not requested with probability (1-ρ). If the computing task is requested in the tth time slot, let the task request indicator value
Figure SMS_73
= 1, otherwise
Figure SMS_61
= 0, that is, P(
Figure SMS_65
=1)=1-P(
Figure SMS_69
=0)=ρ，t∈Γ，P(
Figure SMS_72
=1)=ρ means the probability of computing task being requested is ρ, P(
Figure SMS_62
=0)=1-ρ means that the probability that the computing task is not requested is 1-ρ. In this embodiment, there is no buffer available for queuing computing requests, and the application running on the ocean sensing device is a delay-sensitive application that focuses on the task execution time not being greater than the time slot length, that is,
Figure SMS_66
.

获取第t个时隙时海洋传感设备访问雾节点的信道功率

Figure SMS_74
，其中，
Figure SMS_75
∽
Figure SMS_76
(x)且t∈ Γ，
Figure SMS_77
(x)是
Figure SMS_78
的累积分布函数（CDF，cumulativedistribution function）。Get the channel power of the ocean sensor device accessing the fog node at the tth time slot
Figure SMS_74
,in,
Figure SMS_75
∽
Figure SMS_76
(x) and t∈ Γ,
Figure SMS_77
(x) Yes
Figure SMS_78
The cumulative distribution function (CDF) of .

在获取上述海洋传感设备的实时信息后，执行步骤S2，根据获取的海洋传感设备的实时信息，构建海洋传感设备在该时隙内执行计算任务的计算延迟和能量消耗，以及海洋传感设备在该时隙内将计算任务卸载至雾节点的传输时延和能量消耗。After acquiring the real-time information of the above-mentioned ocean sensor device, execute step S2, and construct the computational delay and energy consumption of the ocean sensor device performing the computational task within the time slot according to the acquired real-time information of the ocean sensor device, as well as the transmission delay and energy consumption of the ocean sensor device offloading the computational task to the fog node within the time slot.

具体的，每个计算任务可以在海洋传感设备上本地执行，也可以卸载到雾节点上由其执行。令

Figure SMS_80
∈{0,1}，j={m,s}，将
Figure SMS_82
作为计算卸载模型的指标，其中，
Figure SMS_85
=1和
Figure SMS_81
=1则分别表示在第t个时隙时，请求的计算任务在海洋传感设备上执行或卸载到雾节点，具体的，当
Figure SMS_83
=1时，则有
Figure SMS_84
=0，表示在第t个时隙时，请求的计算任务在海洋传感设备上执行；当
Figure SMS_86
=1时，则有
Figure SMS_79
=0，表示在第t个时隙时，请求的计算任务卸载到雾节点上执行。因此，计算卸载模型应满足：Specifically, each computing task can be executed locally on the ocean sensor device or offloaded to the fog node for execution.
Figure SMS_80
∈{0,1}, j={m,s},
Figure SMS_82
As an indicator for calculating the unloading model,
Figure SMS_85
=1 and
Figure SMS_81
=1 respectively indicates that at the tth time slot, the requested computing task is executed on the ocean sensor device or offloaded to the fog node. Specifically, when
Figure SMS_83
=1, then
Figure SMS_84
=0, indicating that at the tth time slot, the requested computing task is executed on the ocean sensor device;
Figure SMS_86
=1, then
Figure SMS_79
=0, indicating that at the tth time slot, the requested computing task is offloaded to the fog node for execution. Therefore, the computing offloading model should satisfy:

Figure SMS_87
+
Figure SMS_88
=1，t ∈ Γ
Figure SMS_87
+
Figure SMS_88
=1, t∈Γ

海洋传感设备完成该计算任务需要多个CPU周期的预定频率。将处理1bit大小输入数据所需的CPU周期数记为X，该值可用离线测量的方法根据不同的应用获得，而完成计算任务A(L，

Figure SMS_89
)，则需要W = LX个CPU周期。在第t个时隙内完成计算任务需要W个CPU周期，将该第t个时隙时W个CPU周期的预定频率记为
Figure SMS_90
，即一个时隙对应一个预定频率，该频率可通过DVFS技术调节芯片电压实现。The ocean sensing device needs a predetermined frequency of multiple CPU cycles to complete the computing task. The number of CPU cycles required to process 1 bit of input data is recorded as X, which can be obtained by offline measurement according to different applications, and the number of CPU cycles required to complete the computing task A(L,
Figure SMS_89
), then W = LX CPU cycles are required. It takes W CPU cycles to complete the computing task in the tth time slot, and the predetermined frequency of W CPU cycles in the tth time slot is recorded as
Figure SMS_90
, that is, one time slot corresponds to a predetermined frequency, which can be achieved by adjusting the chip voltage through DVFS technology.

基于在第t个时隙时海洋传感设备完成计算任务所需的W个CPU周期的预定频率

Figure SMS_91
，构建海洋传感设备在t时刻执行计算任务的计算延迟
Figure SMS_92
，为：Based on the predetermined frequency of W CPU cycles required for the ocean sensor device to complete the computing task at the tth time slot
Figure SMS_91
, construct the computational delay of the ocean sensor device performing the computational task at time t
Figure SMS_92
,for:

Figure SMS_93
Figure SMS_93

相应地，构建海洋传感设备在t时刻本地执行计算任务的能量消耗

Figure SMS_94
，为：Accordingly, the energy consumption of constructing ocean sensing equipment to perform computing tasks locally at time t is
Figure SMS_94
,for:

Figure SMS_95
Figure SMS_95

其中，k为有效开关电容，其取值取决于芯片架构。Where k is the effective switch capacitance, and its value depends on the chip architecture.

为了减少移动执行时的计算量，将计算任务A(L，

Figure SMS_96
)卸载到雾节点。在本实施例中，假设计算任务的输入数据量较小，因此反馈的传输延迟也可以忽略。基于海洋传感设备访问雾节点的信道功率
Figure SMS_97
和发射功率
Figure SMS_98
，根据香农公式，构建海洋传感设备在第t个时隙内将计算任务卸载至雾节点的传输时延
Figure SMS_99
，为：In order to reduce the amount of calculation during the move execution, the calculation task A(L,
Figure SMS_96
) is unloaded to the fog node. In this embodiment, it is assumed that the input data volume of the computing task is small, so the transmission delay of the feedback can also be ignored. Channel power of ocean sensing equipment accessing fog nodes
Figure SMS_97
and transmit power
Figure SMS_98
According to Shannon’s formula, the transmission delay of the ocean sensor device to offload the computing task to the fog node in the tth time slot is constructed.
Figure SMS_99
,for:

Figure SMS_100
Figure SMS_100

其中，L表示计算任务的输入数据量，r表示第t个时隙的传输速率，且：Where L represents the input data volume of the computing task, r represents the transmission rate of the tth time slot, and:

Figure SMS_101
Figure SMS_101

上式中，ω表示信道带宽，σ表示信道内部的高斯噪声功率。In the above formula, ω represents the channel bandwidth, and σ represents the Gaussian noise power inside the channel.

相应地，构建海洋传感设备在第t个时隙内将计算任务卸载至雾节点的能量消耗

Figure SMS_102
，为：Accordingly, the energy consumption of constructing the ocean sensor device to offload the computing task to the fog node in the tth time slot is
Figure SMS_102
,for:

Figure SMS_103
。
Figure SMS_103
.

上述步骤S3中，在通过步骤S2获取海洋传感设备在不同情况下执行计算任务的时延和能耗的基础上，首先构建计算任务执行成本模型、海洋传感设备能耗模型和能量收集模型。In the above step S3, based on the delay and energy consumption of the ocean sensor equipment in performing computing tasks under different circumstances obtained in step S2, a computing task execution cost model, an ocean sensor equipment energy consumption model and an energy collection model are first constructed.

在应用能量收集技术的设备中，计算卸载策略的设计比传统的移动云计算系统更加复杂，电池能量水平与时间相关，使得系统决策在不同时隙耦合。因此，最优的计算卸载策略应该在当前计算任务和未来计算任务的计算性能之间取得良好的平衡。In devices that apply energy harvesting technology, the design of computation offloading strategies is more complicated than in traditional mobile cloud computing systems. Battery energy levels are time-dependent, making system decisions coupled at different time slots. Therefore, the optimal computation offloading strategy should achieve a good balance between the computational performance of current and future computational tasks.

任务执行延迟是用户QoS（Quality ofService，服务质量）的关键指标之一，本实施例用QoS来计算卸载策略。由于收集到的能量具有间歇性和散发性，一些请求的计算任务可能无法执行而不得不放弃，例如由于本地计算能量不足，而从海洋传感设备到雾节点的无线通道处于深度衰落，即任务无法卸载。考虑到这一点，对每个被放弃的计算任务按一个成本单位进行惩罚。因此，本实施例构建计算任务执行成本模型，将该执行成本定义为执行延迟和任务丢弃成本的加权和，可以用下式表示:Task execution delay is one of the key indicators of user QoS (Quality of Service). This embodiment uses QoS to calculate the unloading strategy. Since the collected energy is intermittent and scattered, some requested computing tasks may not be executed and have to be abandoned. For example, due to insufficient local computing energy, the wireless channel from the ocean sensor device to the fog node is in deep fading, that is, the task cannot be unloaded. Taking this into account, each abandoned computing task is penalized at a cost unit. Therefore, this embodiment constructs a computing task execution cost model, and defines the execution cost as the weighted sum of the execution delay and the task abandonment cost, which can be expressed as the following formula:

Figure SMS_104
Figure SMS_104

其中，

Figure SMS_105
表示获取到计算任务后却未执行的惩罚，实际上，该惩罚也属于任务执行成本。in,
Figure SMS_105
It indicates the penalty for not executing the computing task after obtaining it. In fact, this penalty is also part of the task execution cost.

如图3所示，能量收集过程可以认为是连续的能量包到达海洋传感设备中，具体的，通过换能器和转换电路将风能、潮汐能、太阳能转换为电能，电能直接或通过电源管理电路间接存储到能量存储设备（如本实施例所述的海洋传感设备）中。通过能量收集，在第t个时隙开始时到达海洋传感设备的能量

Figure SMS_106
，其中，不同时隙的
Figure SMS_107
的最大值为
Figure SMS_108
。在每个时隙中，将基于能量收集技术到达海洋传感设备的部分能量记为
Figure SMS_109
，构建能量收集模型，为：As shown in FIG3 , the energy collection process can be considered as continuous energy packets arriving at the ocean sensor device. Specifically, wind energy, tidal energy, and solar energy are converted into electrical energy through transducers and conversion circuits, and the electrical energy is directly or indirectly stored in the energy storage device (such as the ocean sensor device described in this embodiment) through power management circuits. Through energy collection, the energy arriving at the ocean sensor device at the beginning of the tth time slot
Figure SMS_106
, where different time slots
Figure SMS_107
The maximum value of
Figure SMS_108
In each time slot, the portion of energy reaching the ocean sensing device based on energy harvesting technology is recorded as
Figure SMS_109
, construct the energy harvesting model as:

Figure SMS_110
Figure SMS_110

其中，

Figure SMS_111
，
Figure SMS_112
为所有时隙中到达能量的最大值。in,
Figure SMS_111
,
Figure SMS_112
is the maximum value of the energy arriving in all time slots.

上述能量被收集并存储在海洋传感设备的电池中，在下一时隙中，可以用于本地计算或进行计算卸载。The above energy is collected and stored in the battery of the ocean sensing device, and can be used for local computing or computation offloading in the next time slot.

进而，构建海洋传感设备能耗模型，即海洋传感设备在第t个时隙内所消耗的能量，为：Then, the energy consumption model of the ocean sensing equipment is constructed, that is, the energy consumed by the ocean sensing equipment in the tth time slot is:

Figure SMS_113
Figure SMS_113

同时满足以下能量关系约束：At the same time, the following energy relationship constraints are met:

Figure SMS_114
Figure SMS_114

其中，B表示海洋传感设备的电池能量水平，

Figure SMS_115
表示第t个时隙（即当前时隙）海洋传感设备的电池能量水平。Where, B represents the battery energy level of the ocean sensing device,
Figure SMS_115
Represents the battery energy level of the ocean sensing device at the tth time slot (i.e., the current time slot).

由此，海洋传感设备电池的能量水平变化可以描述为：Therefore, the energy level change of the battery of the ocean sensing device can be described as:

Figure SMS_116
。
Figure SMS_116
.

然后，以长期平均执行成本最小为目标函数，结合所搭建的模型，构建计算卸载模型。Then, taking the minimization of the long-term average execution cost as the objective function and combining the constructed model, a computational offloading model is constructed.

具体的，将优化问题

Figure SMS_117
描述为以长期平均执行成本最小为目标函数，构建计算卸载模型，为：Specifically, the optimization problem
Figure SMS_117
It is described as taking the minimization of the long-term average execution cost as the objective function to build a computation offloading model, which is:

Figure SMS_118
Figure SMS_118

s.t.s.t.

Figure SMS_119
Figure SMS_119

Figure SMS_120
Figure SMS_120

Figure SMS_121
Figure SMS_121

Figure SMS_122
Figure SMS_122

Figure SMS_123
Figure SMS_123

Figure SMS_124
Figure SMS_124

上式中，

Figure SMS_125
表示第t个时隙时的最大传输功率（即最大发射功率），
Figure SMS_126
表示第t个时隙时的最大CPU预定频率，T表示时隙的总个数，t=0,1,...,T-1。In the above formula,
Figure SMS_125
represents the maximum transmission power (i.e., maximum transmit power) at the tth time slot,
Figure SMS_126
It represents the maximum CPU scheduled frequency in the tth time slot, T represents the total number of time slots, t=0,1,...,T-1.

在构建完成计算卸载模型后，执行步骤S4，即根据所获取的实时信息，利用动态算法求解计算卸载模型，求解获得当前时隙的最优计算卸载决策。After the computation offloading model is constructed, step S4 is executed, that is, according to the acquired real-time information, the computation offloading model is solved by using a dynamic algorithm to obtain the optimal computation offloading decision for the current time slot.

在所考虑的卸载系统中，系统状态由任务请求、可收集能量、电池能量水平和信道状态组成，动作为能量收集和计算卸载决策，包括CPU周期调度频率和分配的发射功率。操作仅依赖于当前系统状态。此外，由于是以长期平均执行成本最小为目标函数，因此，问题

Figure SMS_127
实际上是一个马尔科夫决策过程（MDP，MarkovDecision Process）问题。原则上，问题
Figure SMS_128
可以通过标准的MDP算法得到最优解，如相对值迭代算法和线性规划重公式方法。然而，这两种算法均需要使用有限状态来描述系统，并离散可行行动集。由于MDP算法是基于数值迭代的，因此很难获得解决方案。此外，量化状态和可行动作集可能会导致严重的性能退化，海洋传感设备资源十分有限，无法满足存储最优策略的内存需求。因此，本实施例提出一个动态求解算法来解决问题
Figure SMS_129
。In the offloading system considered, the system state consists of task requests, harvestable energy, battery energy level, and channel status, and the actions are energy harvesting and computation offloading decisions, including CPU cycle scheduling frequency and allocated transmit power. The operation depends only on the current system state. In addition, since the objective function is to minimize the long-term average execution cost, the problem
Figure SMS_127
It is actually a Markov decision process (MDP) problem. In principle, the problem
Figure SMS_128
The optimal solution can be obtained through standard MDP algorithms, such as the relative value iteration algorithm and the linear programming reformulation method. However, both algorithms require the use of finite states to describe the system and discretize the feasible action set. Since the MDP algorithm is based on numerical iteration, it is difficult to obtain a solution. In addition, quantizing the state and the feasible action set may cause serious performance degradation. The resources of marine sensor equipment are very limited and cannot meet the memory requirements for storing the optimal strategy. Therefore, this embodiment proposes a dynamic solution algorithm to solve the problem
Figure SMS_129
.

计算卸载决策（即计算卸载模型的解）在某一个时间节点可以用一个多维向量描述，李雅普诺夫函数（Lyapunov函数）则是对这个多为向量所表示的状态的非负、标量的表述。若系统朝向不被期望的方向发展，那么Lyapunov函数就会变大，因此，将Lyapunov函数沿着x轴的负方向逼近到0，使之趋于稳定，当达到Lyapunov稳定后，则认为此时的计算卸载决策（即由多维向量构成的解）为当前时间节点的最优决策。The computational offloading decision (i.e., the solution of the computational offloading model) can be described by a multidimensional vector at a certain time node, and the Lyapunov function is a non-negative, scalar representation of the state represented by this multidimensional vector. If the system develops in an undesirable direction, the Lyapunov function will become larger. Therefore, the Lyapunov function is approached to 0 along the negative direction of the x-axis to make it stable. When Lyapunov stability is reached, the computational offloading decision (i.e., the solution composed of multidimensional vectors) at this time is considered to be the optimal decision at the current time node.

在本实施例中，首先，基于当前时隙的海洋传感设备的电池能量水平，构建李雅普诺夫函数，具体如下式所示：In this embodiment, firstly, based on the battery energy level of the ocean sensing device in the current time slot, a Lyapunov function is constructed, as shown in the following formula:

Figure SMS_130
Figure SMS_130

然后，构建李雅普诺夫漂移函数，具体如下式所示：Then, construct the Lyapunov drift function, as shown in the following formula:

Figure SMS_131
Figure SMS_131

在李雅普诺夫漂移函数中引入期望，引入的期望如下式所示：Introducing expectation into the Lyapunov drift function, the introduced expectation is shown as follows:

Figure SMS_132
Figure SMS_132

上式中，

Figure SMS_133
表示取期望。In the above formula,
Figure SMS_133
Indicates expectation.

自然地，Naturally,

Figure SMS_134
Figure SMS_134

其中，C是大于0的有上界的常数。Where C is a constant with an upper bound greater than 0.

由于求解问题

Figure SMS_135
（即计算卸载模型的目标函数）获得最优解，其计算复杂、难以计算，因此通过上述方式适当放大形成问题
Figure SMS_136
，通过求解问题
Figure SMS_137
得到问题
Figure SMS_138
的最优解集。所形成的问题
Figure SMS_139
为：Because of the problem
Figure SMS_135
(i.e., calculating the objective function of the unloading model) to obtain the optimal solution, the calculation is complex and difficult to calculate, so the above method is appropriately enlarged to form a problem
Figure SMS_136
, by solving the problem
Figure SMS_137
Get the question
Figure SMS_138
The optimal solution set of . The problem formed
Figure SMS_139
for:

Figure SMS_140
Figure SMS_140

s.t.s.t.

Figure SMS_141
Figure SMS_141

在问题

Figure SMS_142
的基础上再次化简，即对所有时隙的李雅普诺夫漂移函数进行累加，得：In question
Figure SMS_142
Based on this, we can simplify it again, that is, accumulate the Lyapunov drift functions of all time slots, and get:

Figure SMS_143
Figure SMS_143

上式中，

Figure SMS_144
表示第0个时隙时海洋传感设备的电池能量水平，即海洋传感设备的初始电池能量水平，通常设
Figure SMS_145
=0；
Figure SMS_146
表示在经过T个时隙（指0~T-1共T个时隙）后，第T个时隙时海洋传感设备的电池能量水平。In the above formula,
Figure SMS_144
It represents the battery energy level of the ocean sensor device at the 0th time slot, that is, the initial battery energy level of the ocean sensor device.
Figure SMS_145
=0;
Figure SMS_146
It indicates the battery energy level of the ocean sensor device at the Tth time slot after T time slots (a total of T time slots from 0 to T-1).

对上述等式两边取期望，得：Taking the expectation on both sides of the above equation, we get:

Figure SMS_147
Figure SMS_147

为了使上式成立，令

Figure SMS_148
，则可得到问题
Figure SMS_149
，为：In order to make the above equation valid, let
Figure SMS_148
, then we can get the problem
Figure SMS_149
,for:

Figure SMS_150
Figure SMS_150

s.t.s.t.

Figure SMS_151
Figure SMS_151

因此，李雅普诺夫漂移函数和李雅普诺夫漂移加惩罚函数可以表示为：Therefore, the Lyapunov drift function and the Lyapunov drift plus penalty function can be expressed as:

Figure SMS_152
Figure SMS_152

在每个时隙内求其最小值，借助权重V来调节对

Figure SMS_153
和
Figure SMS_154
的重视程度，因此形成新的问题
Figure SMS_155
，即：Find the minimum value in each time slot and adjust the
Figure SMS_153
and
Figure SMS_154
The level of attention paid to this issue has led to new problems
Figure SMS_155
,Right now:

Figure SMS_156
Figure SMS_156

s.t.s.t.

Figure SMS_157
Figure SMS_157

通过上述方案，利用李雅普诺夫漂移函数和李雅普诺夫漂移加惩罚函数，化简计算卸载模型，将难以求解最优解的问题

Figure SMS_158
转换值问题
Figure SMS_159
，通过求解问题
Figure SMS_160
获取与最优解相差无几且包含最优解的解集。Through the above scheme, using Lyapunov drift function and Lyapunov drift plus penalty function, the calculation of unloading model is simplified, and the problem of finding the optimal solution is difficult to solve.
Figure SMS_158
Conversion value problem
Figure SMS_159
, by solving the problem
Figure SMS_160
Get a set of solutions that are close to the optimal solution and include the optimal solution.

在完成计算卸载模型的化简后，通过下述算法求解问题

Figure SMS_161
的渐进最优解，即求解当前时隙内的最优计算卸载决策。After completing the simplification of the computational unloading model, the problem is solved by the following algorithm:
Figure SMS_161
The asymptotically optimal solution of , that is, solving the optimal computation offloading decision in the current time slot.

首先，在第t个时隙开始时，获取当前时隙海洋传感设备所请求计算任务的任务请求指示值

Figure SMS_162
（任务请求指示值
Figure SMS_163
的取值为{0，1}，代表是否获取到计算任务）、海洋传感设备的电池能量水平
Figure SMS_164
和海洋传感设备访问雾节点的信道功率
Figure SMS_165
。First, at the beginning of the tth time slot, obtain the task request indication value of the computing task requested by the ocean sensing device in the current time slot
Figure SMS_162
(Task request indication value
Figure SMS_163
The value of is {0, 1}, indicating whether the computing task is obtained), the battery energy level of the ocean sensing device
Figure SMS_164
and channel power for ocean sensor devices to access fog nodes
Figure SMS_165
.

其次，根据上述获取的当前时隙的信息，代入化简后的计算卸载模型的目标函数中进行求解，求解得到由多维向量构成的解，即确定计算卸载模型的指标

Figure SMS_166
、CPU周期的预定频率
Figure SMS_167
、海洋传感设备访问雾节点的发射功率
Figure SMS_168
基于能量收集技术到达海洋传感设备的部分能量
Figure SMS_169
。其中，化简后的计算卸载模型如下式所示：Secondly, according to the information of the current time slot obtained above, it is substituted into the objective function of the simplified computation offloading model for solving, and the solution consisting of multi-dimensional vectors is obtained, that is, the index of the computation offloading model is determined.
Figure SMS_166
, the predetermined frequency of CPU cycles
Figure SMS_167
, Transmission power of marine sensor equipment accessing fog nodes
Figure SMS_168
Part of the energy reaching ocean sensing devices based on energy harvesting technology
Figure SMS_169
Among them, the simplified calculation offloading model is shown as follows:

Figure SMS_170
Figure SMS_170

s.t.s.t.

Figure SMS_171
Figure SMS_171

之后，根据计算确定的结果更新海洋传感设备的电池能量队列，即更新

Figure SMS_172
，并判断是否达到当前时隙的稳定值，即判断是否满足计算卸载模型的约束条件，若满足，则认为达到当前时隙的稳定值，将计算确定的结果作为当前时隙内的最优计算卸载决策输出；否则计算下一时隙的最优计算卸载决策。Afterwards, the battery energy queue of the ocean sensor device is updated according to the calculated result, that is, the battery energy queue of the ocean sensor device is updated.
Figure SMS_172
, and judge whether the stable value of the current time slot is reached, that is, judge whether the constraints of the computational offloading model are met. If so, it is considered that the stable value of the current time slot is reached, and the result determined by the calculation is output as the optimal computational offloading decision in the current time slot; otherwise, the optimal computational offloading decision for the next time slot is calculated.

考虑到海洋传感设备的资源（计算、存储和能源）及海洋节点动态拓扑的限制，传统卸载方法无法适用，因此，本实施例提供了一种基于李雅普诺夫（Lyapunov）优化的动态计算卸载策略，该策略复杂度低，适配海洋节点的弱计算能力，充分利用各海洋节点的能源和计算能力，对计算任务进行更加合理的分配，从而达到减小平均决策时延和系统能耗的目的，保证服务质量QoS。Considering the limitations of the resources (computing, storage and energy) of marine sensor equipment and the dynamic topology of marine nodes, traditional offloading methods are not applicable. Therefore, this embodiment provides a dynamic computing offloading strategy based on Lyapunov optimization. This strategy has low complexity, adapts to the weak computing power of marine nodes, makes full use of the energy and computing power of each marine node, and allocates computing tasks more reasonably, thereby achieving the purpose of reducing the average decision delay and system energy consumption, and ensuring the quality of service QoS.

实施例二Embodiment 2

本实施例提供了一种基于能量收集技术的海洋传感网络计算卸载系统，包括：This embodiment provides a marine sensor network computing unloading system based on energy harvesting technology, including:

信息获取模块，用于在每一时隙内，获取当前时隙海洋传感设备的实时信息，所述实时信息包括海洋传感设备所请求计算任务的任务请求指示值、海洋传感设备访问雾节点的信道功率以及海洋传感设备的电池能量水平；An information acquisition module is used to acquire real-time information of the ocean sensor device in the current time slot in each time slot, wherein the real-time information includes a task request indication value of the computing task requested by the ocean sensor device, a channel power for the ocean sensor device to access the fog node, and a battery energy level of the ocean sensor device;

模型构建模块，用于构建海洋传感设备在当前时隙内执行计算任务的计算延迟和能量消耗，以及将计算任务卸载至雾节点的传输时延和能量消耗，搭建计算任务执行成本模型、海洋传感设备能耗模型和能量收集模型，以长期平均执行成本最小为目标函数，结合所搭建的模型，构建计算卸载模型；The model building module is used to construct the computational delay and energy consumption of the ocean sensor equipment in executing the computational task in the current time slot, as well as the transmission delay and energy consumption of offloading the computational task to the fog node, build the computational task execution cost model, the ocean sensor equipment energy consumption model and the energy collection model, and build the computational offloading model with the minimum long-term average execution cost as the objective function in combination with the built model;

卸载决策求解模块，用于根据所获取的实时信息，利用动态算法求解计算卸载模型，求解获得当前时隙的最优计算卸载决策。The offloading decision solving module is used to solve the computing offloading model using a dynamic algorithm based on the acquired real-time information, and to obtain the optimal computing offloading decision for the current time slot.

以上实施例二中涉及的各步骤与方法实施例一相对应，具体实施方式可参见实施例一的相关说明部分。The steps involved in the above embodiment 2 correspond to those in the method embodiment 1. For the specific implementation method, please refer to the relevant description part of the embodiment 1.

本领域技术人员应该明白，上述本发明的各模块或各步骤可以用通用的计算机装置来实现，可选地，它们可以用计算装置可执行的程序代码来实现，从而，可以将它们存储在存储装置中由计算装置来执行，或者将它们分别制作成各个集成电路模块，或者将它们中的多个模块或步骤制作成单个集成电路模块来实现。本发明不限制于任何特定的硬件和软件的结合。Those skilled in the art should understand that the modules or steps of the present invention described above can be implemented by a general-purpose computer device, or alternatively, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. The present invention is not limited to any specific combination of hardware and software.

以上所述仅为本发明的优选实施例而已，并不用于限制本发明，对于本领域的技术人员来说，本发明可以有各种更改和变化。凡在本发明的精神和原则之内，所作的任何修改、等同替换、改进等，均应包含在本发明的保护范围之内。The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

上述虽然结合附图对本发明的具体实施方式进行了描述，但并非对本发明保护范围的限制，所属领域技术人员应该明白，在本发明的技术方案的基础上，本领域技术人员不需要付出创造性劳动即可做出的各种修改或变形仍在本发明的保护范围以内。Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Technical personnel in the relevant field should understand that various modifications or variations that can be made by technical personnel in the field without creative work on the basis of the technical solution of the present invention are still within the scope of protection of the present invention.

Claims (5)

1. The ocean sensing network calculation unloading method based on the energy collection technology is characterized by comprising the following steps of:
acquiring real-time information of ocean sensing equipment in a current time slot in each time slot; the real-time information comprises a task request indicated value of a calculation task requested by the marine sensing equipment, channel power of the marine sensing equipment for accessing the fog node and battery energy level of the marine sensing equipment;
constructing a calculation delay and energy consumption of the marine sensing device for executing the calculation task in the current time slot, and a transmission delay and energy consumption for unloading the calculation task to the fog node;
constructing a calculation task execution cost model, an ocean sensing equipment energy consumption model and an energy collection model, taking the minimum long-term average execution cost as an objective function, and constructing a calculation unloading model by combining the constructed models;
according to the acquired real-time information, a dynamic algorithm is utilized to solve a calculation unloading model, and an optimal calculation unloading decision of the current time slot is obtained;
the computational offload model is:
Figure QLYQS_1
s.t.
Figure QLYQS_2
Figure QLYQS_3
Figure QLYQS_4
Figure QLYQS_5
Figure QLYQS_6
Figure QLYQS_7
wherein ,
Figure QLYQS_8
，
Figure QLYQS_11
computing task execution cost representing the t-th slot, < >>
Figure QLYQS_13
Indicating penalty not performed after acquisition of the computational task,/->
Figure QLYQS_10
A calculation delay representing the local execution of a calculation task by the marine sensor device in the t-th time slot, a->
Figure QLYQS_12
Representing the transmission delay of the marine sensing device for offloading the computation task to the mist node in the t-th time slot, a +.>
Figure QLYQS_14
Index representing computational offload model, +.>
Figure QLYQS_15
=1 and
Figure QLYQS_9
=1 means that at time slot t, the requested computation task is performed on the marine sensing device and the computation task is offloaded to the fog node;
Figure QLYQS_16
，
Figure QLYQS_17
representing the energy consumed by the marine sensing device in the t-th time slot,/for>
Figure QLYQS_18
Representing the battery energy level of the marine sensing device at time t, +.>
Figure QLYQS_19
Energy consumption representing the local execution of a computational task by a marine sensor device in the t-th time slot, a->
Figure QLYQS_20
Representing energy consumption of the marine sensing device to offload a computing task to a fog node in a t-th time slot;
Figure QLYQS_21
indicating maximum transmission power at the t-th slot,/->
Figure QLYQS_22
A predetermined frequency representing the maximum CPU period at the T-th slot, t=0, 1..t-1, T represents the total number of time slots, +.>
Figure QLYQS_23
Indicating the desire to get->
Figure QLYQS_24
A predetermined frequency representing the CPU period at the current t-th slot,/or->
Figure QLYQS_25
Indicating the transmit power of the marine sensing device accessing the mist node,/->
Figure QLYQS_26
Representing a portion of the energy reaching the marine sensing device,
Figure QLYQS_27
representing the scheduled frequency of w CPU cycles required by the ocean sensing equipment to complete the calculation task at the t time slot;
solving an optimal computational offload decision in a current time slot using a dynamic algorithm, comprising:
constructing a Lyapunov function;
constructing a Lyapunov drift function and a Lyapunov drift plus penalty function according to the Lyapunov function;
simplifying and calculating an unloading model according to the Lyapunov drift function and the Lyapunov drift plus penalty function;
solving the simplified calculation unloading model to obtain an optimal calculation unloading decision in the current time slot;
the solving includes:
when the t time slot starts, acquiring a task request instruction value of a calculation task requested by the ocean sensing equipment in the current time slot
Figure QLYQS_28
Battery energy level of marine sensor device +.>
Figure QLYQS_29
And channel power of marine sensing device access fog node +.>
Figure QLYQS_30
；
Substituting the information of the current time slot into the simplified calculation unloading model to solve, and determining the index of the calculation unloading model of the current time slot
Figure QLYQS_31
Predetermined frequency of CPU cycle->
Figure QLYQS_32
Transmitting power of ocean sensing equipment for accessing fog node
Figure QLYQS_33
Partial energy reaching the marine sensor system based on energy harvesting technology>
Figure QLYQS_34
；
And updating a battery energy queue of the ocean sensing equipment according to the result of calculation determination, judging whether a stable value of the current time slot is reached, namely judging whether constraint conditions of a calculation unloading model are met, if so, outputting the result of calculation determination as an optimal calculation unloading decision in the current time slot, otherwise, calculating an optimal calculation unloading decision of the next time slot.
2. The energy harvesting technique-based ocean sensing network computing offload of claim 1, wherein the predetermined frequency based on w CPU cycles required by the ocean sensing device to complete the computing task at the t-th time slot
Figure QLYQS_35
Constructing a computation delay for the marine sensor device to perform the computation task locally in the t-th time slot +.>
Figure QLYQS_36
And energy consumption->
Figure QLYQS_37
The formula is:
Figure QLYQS_38
Figure QLYQS_39
wherein ,kin order to be able to switch the capacitance in an efficient way,wthe number of CPU cycles is represented by w=1, …, W.
3. The energy harvesting technique-based ocean sensing network computing offloading method of claim 1, wherein the channel power of the mist node is accessed based on ocean sensing equipment
Figure QLYQS_40
And transmit power->
Figure QLYQS_41
According to shannon formula, constructing transmission delay +.f for the marine sensing device to offload computing task to fog node in the t-th time slot>
Figure QLYQS_42
And energy consumption->
Figure QLYQS_43
The formula is:
Figure QLYQS_44
Figure QLYQS_45
wherein L represents the input data amount of the calculation task, r represents the transmission rate of the t-th time slot, and
Figure QLYQS_46
in the above equation, ω denotes a channel bandwidth, and σ denotes gaussian noise power inside the channel.
4. The energy harvesting technique-based ocean sensing network computing offload method of claim 1, wherein the fraction of energy arriving at the ocean sensing apparatus during each time slot based on the energy harvesting technique is recorded as
Figure QLYQS_47
An energy collection model is constructed as follows:
Figure QLYQS_48
wherein ,
Figure QLYQS_49
representing the energy reaching the marine sensing device at the beginning of the t-th time slot.
5. An ocean sensing network computing and unloading system based on an energy collection technology is characterized by comprising:
the information acquisition module is used for acquiring real-time information of the marine sensing equipment in the current time slot in each time slot, wherein the real-time information comprises a task request indication value of a calculation task requested by the marine sensing equipment, channel power of the marine sensing equipment for accessing the fog node and battery energy level of the marine sensing equipment;
the model construction module is used for constructing calculation delay and energy consumption of a calculation task executed by the marine sensing equipment in the current time slot, transmission delay and energy consumption of unloading the calculation task to the fog node, constructing a calculation task execution cost model, a marine sensing equipment energy consumption model and an energy collection model, taking the minimum long-term average execution cost as an objective function, and constructing a calculation unloading model by combining the constructed model;
the unloading decision solving module is used for solving a calculation unloading model by utilizing a dynamic algorithm according to the acquired real-time information, and solving and obtaining an optimal calculation unloading decision of the current time slot;
the computational offload model is:
Figure QLYQS_50
s.t.
Figure QLYQS_51
Figure QLYQS_52
Figure QLYQS_53
Figure QLYQS_54
Figure QLYQS_55
Figure QLYQS_56
wherein ,
Figure QLYQS_59
，
Figure QLYQS_61
computing task execution cost representing the t-th slot, < >>
Figure QLYQS_63
Indicating penalty not performed after acquisition of the computational task,/->
Figure QLYQS_58
A calculation delay representing the local execution of a calculation task by the marine sensor device in the t-th time slot, a->
Figure QLYQS_60
Indicating that the marine sensing device is offloading computing tasks to fog in the t-th time slotTransmission delay of node->
Figure QLYQS_62
Index representing computational offload model, +.>
Figure QLYQS_64
=1 and
Figure QLYQS_57
=1 means that at time slot t, the requested computation task is performed on the marine sensing device and the computation task is offloaded to the fog node;
Figure QLYQS_65
，
Figure QLYQS_66
representing the energy consumed by the marine sensing device in the t-th time slot,/for>
Figure QLYQS_67
Representing the battery energy level of the marine sensing device at time t, +.>
Figure QLYQS_68
Energy consumption representing the local execution of a computational task by a marine sensor device in the t-th time slot, a->
Figure QLYQS_69
Representing energy consumption of the marine sensing device to offload a computing task to a fog node in a t-th time slot;
Figure QLYQS_70
indicating maximum transmission power at the t-th slot,/->
Figure QLYQS_71
A predetermined frequency representing the maximum CPU period at time slot T, t=0, 1,..,t represents the total number of time slots, +.>
Figure QLYQS_72
Indicating the desire to get->
Figure QLYQS_73
A predetermined frequency representing the CPU period at the current t-th slot,/or->
Figure QLYQS_74
Indicating the transmit power of the marine sensing device accessing the mist node,/->
Figure QLYQS_75
Representing part of the energy reaching the marine sensing device, < +.>
Figure QLYQS_76
Representing the scheduled frequency of w CPU cycles required by the ocean sensing equipment to complete the calculation task at the t time slot;
solving an optimal computational offload decision in a current time slot using a dynamic algorithm, comprising:
constructing a Lyapunov function;
constructing a Lyapunov drift function and a Lyapunov drift plus penalty function according to the Lyapunov function;
simplifying and calculating an unloading model according to the Lyapunov drift function and the Lyapunov drift plus penalty function;
solving the simplified calculation unloading model to obtain an optimal calculation unloading decision in the current time slot;
the solving includes:
when the t time slot starts, acquiring a task request instruction value of a calculation task requested by the ocean sensing equipment in the current time slot
Figure QLYQS_77
Battery energy level of marine sensor device +.>
Figure QLYQS_78
And channel power of marine sensing device access fog node +.>
Figure QLYQS_79
；
Substituting the information of the current time slot into the simplified calculation unloading model to solve, and determining the index of the calculation unloading model of the current time slot
Figure QLYQS_80
Predetermined frequency of CPU cycle->
Figure QLYQS_81
Transmitting power of ocean sensing equipment for accessing fog node
Figure QLYQS_82
Partial energy reaching the marine sensor system based on energy harvesting technology>
Figure QLYQS_83
；
And updating a battery energy queue of the ocean sensing equipment according to the result of calculation determination, judging whether a stable value of the current time slot is reached, namely judging whether constraint conditions of a calculation unloading model are met, if so, outputting the result of calculation determination as an optimal calculation unloading decision in the current time slot, otherwise, calculating an optimal calculation unloading decision of the next time slot.
CN202310152246.XA 2023-02-23 2023-02-23 Ocean sensing network calculation unloading method and system based on energy collection technology Active CN115843070B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310152246.XA CN115843070B (en) 2023-02-23 2023-02-23 Ocean sensing network calculation unloading method and system based on energy collection technology

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310152246.XA CN115843070B (en) 2023-02-23 2023-02-23 Ocean sensing network calculation unloading method and system based on energy collection technology

Publications (2)

Publication Number Publication Date
CN115843070A CN115843070A (en) 2023-03-24
CN115843070B true CN115843070B (en) 2023-06-16

Family

ID=85580049

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310152246.XA Active CN115843070B (en) 2023-02-23 2023-02-23 Ocean sensing network calculation unloading method and system based on energy collection technology

Country Status (1)

Country Link
CN (1) CN115843070B (en)

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112383931A (en) * 2020-11-12 2021-02-19 东华大学 Method for optimizing cost and time delay in multi-user mobile edge computing system

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10111024B2 (en) * 2015-07-10 2018-10-23 Lg Electronics Inc. Method and apparatus for an input data processing via a local computing or offloading based on power harvesting in a wireless communication system
CN109829332B (en) * 2019-01-03 2021-05-04 武汉理工大学 A joint computing offloading method and device based on energy harvesting technology
CN113114733B (en) * 2021-03-24 2022-07-08 重庆邮电大学 Distributed task unloading and computing resource management method based on energy collection
CN114650568B (en) * 2022-03-18 2024-09-06 南京徐庄科技创业服务中心有限公司 Distributed unloading method based on energy collection in mobile Ad Hoc cloud
CN114385272B (en) * 2022-03-24 2022-07-05 山东省计算中心（国家超级计算济南中心） Ocean task oriented online adaptive computing unloading method and system
CN115696451A (en) * 2022-10-10 2023-02-03 广州大学 Optimization method for energy and task scheduling of edge computing system
CN115686669B (en) * 2022-10-17 2023-05-23 中国矿业大学 Intelligent calculation unloading method for mine Internet of things assisted by energy collection

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112383931A (en) * 2020-11-12 2021-02-19 东华大学 Method for optimizing cost and time delay in multi-user mobile edge computing system

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
Energy-Efficient Data Collection Scheme Based on Mobile Edge Computing in WSNs;Xin Li等;《IEEE》;全文 *
基于云雾混合计算的车联网联合资源分配算法;唐伦;肖娇;魏延南;赵国繁;陈前斌;;电子与信息学报(08);全文 *

Also Published As

Publication number Publication date
CN115843070A (en) 2023-03-24

Similar Documents

Publication Publication Date Title
Chen et al. QoS-aware robotic streaming workflow allocation in cloud robotics systems
CN112004239A (en) Computing unloading method and system based on cloud edge cooperation
CN113452956B (en) Intelligent distribution method and system for power transmission line inspection tasks
TWI542986B (en) Management system, multi-core system and its management method
CN114285853B (en) Task unloading method based on end edge cloud cooperation in equipment-intensive industrial Internet of things
CN112835715B (en) Method and device for determining unmanned aerial vehicle task offloading strategy based on reinforcement learning
Crutcher et al. Hyperprofile-based computation offloading for mobile edge networks
CN117858109A (en) User association, task unloading and resource allocation optimization method based on digital twin
CN114007231A (en) Heterogeneous unmanned aerial vehicle data unloading method and device, electronic equipment and storage medium
Nguyen et al. Modeling multi-constrained fog-cloud environment for task scheduling problem
CN116954866A (en) Edge cloud task scheduling method and system based on deep reinforcement learning
CN113747450A (en) Service deployment method and device in mobile network and electronic equipment
CN109375999A (en) A Bayesian Network-based Random Task Transfer Method for MEC
Zhan Elastic Scheduling of Micro-Modules in Edge Computing Based on LSTM Prediction
CN115843070B (en) Ocean sensing network calculation unloading method and system based on energy collection technology
Gao et al. Markov decision process‐based computation offloading algorithm and resource allocation in time constraint for mobile cloud computing
CN111162852B (en) Ubiquitous power Internet of things access method based on matching learning
CN118316024A (en) Electric power intelligent scheduling method based on artificial intelligence
Belkout et al. A load balancing and routing strategy in fog computing using deep reinforcement learning
CN117573339A (en) Dynamic capacity expansion and contraction method, device and storage medium for satellite edge computing services
CN115913965A (en) Task timeliness-based associated slice resource allocation method
CN113157344B (en) DRL-based energy consumption perception task unloading method in mobile edge computing environment
Wang et al. Offloading strategies for mobile edge computing based on multi-attribute preferences in smart grids
CN114979027A (en) Task offloading and resource allocation method for offshore edge computing network based on energy harvesting
Kallimani et al. Dynamic power management modelling of a wireless sensor node

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant