CN115843070A - Ocean sensor network computing unloading method and system based on energy collection technology - Google Patents
Ocean sensor network computing unloading method and system based on energy collection technology Download PDFInfo
- Publication number
- CN115843070A CN115843070A CN202310152246.XA CN202310152246A CN115843070A CN 115843070 A CN115843070 A CN 115843070A CN 202310152246 A CN202310152246 A CN 202310152246A CN 115843070 A CN115843070 A CN 115843070A
- Authority
- CN
- China
- Prior art keywords
- time slot
- ocean
- task
- calculation
- energy
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 43
- 238000005516 engineering process Methods 0.000 title claims abstract description 31
- 238000004364 calculation method Methods 0.000 claims description 127
- 238000005265 energy consumption Methods 0.000 claims description 37
- 230000005540 biological transmission Effects 0.000 claims description 21
- 238000003306 harvesting Methods 0.000 claims description 17
- 230000007774 longterm Effects 0.000 claims description 12
- 238000009472 formulation Methods 0.000 claims 1
- 239000003595 mist Substances 0.000 claims 1
- 230000006870 function Effects 0.000 description 32
- 230000008569 process Effects 0.000 description 5
- 239000013598 vector Substances 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 125000004432 carbon atom Chemical group C* 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 230000036961 partial effect Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 230000004044 response Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000002829 reductive effect Effects 0.000 description 2
- 238000013468 resource allocation Methods 0.000 description 2
- 238000003860 storage Methods 0.000 description 2
- 239000002699 waste material Substances 0.000 description 2
- 241000282414 Homo sapiens Species 0.000 description 1
- 230000001133 acceleration Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000003990 capacitor Substances 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000001186 cumulative effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000005315 distribution function Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000004146 energy storage Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000011065 in-situ storage Methods 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 230000000670 limiting effect Effects 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000036962 time dependent Effects 0.000 description 1
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 method and a system for calculating and unloading an ocean sensing network based on an energy collection technology, and relates to the technical field of ocean observation sensing networks.
Description
Technical Field
The invention relates to the technical field of ocean observation sensor networks, in particular to an energy collection technology-based ocean sensor network computation unloading method and system.
Background
The traditional ocean observation is mainly sea-based observation mainly based on a survey ship and a submarine buoy or space-based observation based on satellite remote sensing and aviation observation. Due to the complexity and uniqueness of marine environment, the development of marine science is always restricted by the problems of short duration, discontinuity and the like of marine observation data. The seabed observation network derived from the U.S. naval underwater acoustic monitoring system in the cold war period is a third ocean science observation platform established by human beings. Under the promotion of novel technologies such as modern sensors, underwater robots, submarine optical fiber cables, internet of things and big data, the submarine observation network integrates subjects such as physical oceans, ocean chemistry, ocean geophysical, ocean ecology and the like, solves the technical problem of obtaining ocean observation data in real time at high resolution in deep sea and extreme environments, can go deep into the ocean to observe and know the ocean, and realizes all-weather, long-term, continuous, comprehensive, real-time and in-situ observation from the seabed to the sea surface.
The establishment of a distributed, networked, interactive and comprehensive intelligent three-dimensional observation network is a development trend of marine science observation. With the application of the internet of things technology in the ocean field, observation means such as observation stations, observation nodes, satellite remote sensing, unmanned surface vessels and the like which are scattered at all places are integrated through unified and universal data standards to carry out cooperative work, so that a layered, integrated and intelligent air-sky-ocean integrated stereo observation network covering the nearshore, the region and the global sea area is formed. Due to the limited resources (e.g., computing and energy capacity) of the sensing device, the computing power may be greatly reduced when the energy is low, and may even result in a service failure. The computing task in the sensing equipment is unloaded to the remote cloud through the sensing network gateway, so that the computing problem can be effectively solved. However, cloud processing cannot meet the stringent latency requirements of many observation applications, which typically require real-time processing and response (e.g., disaster response).
The fog node is a basic element of the fog architecture, and may be any device that provides computing, network, storage, and acceleration elements of the fog architecture, such as an industrial Controller, a switch, a router, an embedded server, a complex gateway, a Programmable Logic Controller (PLC), and an intelligent internet of things node (e.g., a video surveillance camera). Considering that the fog nodes have the functions of calculating and storing network resources, the service performance of the marine observation network can be effectively improved by deploying the fog nodes near the marine sensing equipment.
The fog node is usually connected to a sensor network gateway and used for processing computing tasks from sensor network equipment so as to provide instant service response; the energy harvesting technology can collect a small amount of unconventional energy easily obtained in the environment and convert the unconventional energy into electric energy to continuously supply power to the sensing equipment. However, in a marine environment with a complex environment and dynamically changing topology, the traditional task allocation method cannot be effectively applied, and due to the requirement of Quality of Service (QoS), the traditional unloading method unloads all computation tasks to the nearest fog nodes, which may cause overload of some fog nodes and insufficient load of sensing equipment; or the sensing equipment undertakes a calculation task, so that the power of the sensing equipment is rapidly reduced, and the task fails. Therefore, for a marine environment in which nodes dynamically change, the existing task allocation/offloading method cannot be applied to a marine sensor network to which an energy collection technology is applied.
Disclosure of Invention
In order to solve the defects of the prior art, the invention provides a method and a system for calculating and unloading an ocean sensing network based on an energy collection technology, which are suitable for the ocean sensing network applying the energy collection technology, provide an optimal calculation and unloading strategy, realize the determination of a proper task unloading proportion for ocean sensing equipment and fog nodes, improve the calculation performance of the ocean sensing network and avoid the waste of node resources.
In a first aspect, the present disclosure provides a marine sensor network computation offloading method based on an energy collection technology.
A marine sensor network computing unloading method based on an energy collection technology comprises the following steps:
in each time slot, acquiring real-time information of the current time slot marine sensing equipment; the real-time information comprises a task request indicating value of a calculation task requested by the ocean sensing equipment, channel power of an access fog node of the ocean sensing equipment and battery energy level of the ocean sensing equipment;
constructing the calculation delay and energy consumption of the ocean sensing equipment for executing the calculation task in the current time slot, and the transmission delay and energy consumption for unloading the calculation task to the fog node;
building a calculation task execution cost model, an ocean sensing equipment energy consumption model and an energy collection model, and building a calculation unloading model by taking the minimum long-term average execution cost as a target function and combining the built model;
and solving the calculation unloading model by using a dynamic algorithm according to the acquired real-time information, and solving to acquire the optimal calculation unloading decision of the current time slot.
In a further technical scheme, the preset frequency of W CPU cycles required by the marine sensing equipment to complete the calculation task at the t-th time slot is based on
And constructing the calculation delay of the marine sensing equipment for locally executing the calculation task in the t time slot
And energy expenditure->
The formula is as follows:
wherein ,kin order to effectively switch the capacitance on and off,windicates the number of CPU cycles, W =1, …, W.
The further technical scheme is that the channel power of the fog nodes is accessed based on the ocean sensing equipment
And an emission power->
Constructing the transmission time delay for the ocean sensing equipment to unload the calculation task to the fog node in the t-th time slot according to the Shannon formula
And energy expenditure->
The formula is as follows:
wherein L represents the input data amount of the calculation task, r represents the transmission rate of the t-th time slot, and
in the above equation, ω represents a channel bandwidth and σ represents a gaussian noise power inside the channel.
According to a further technical scheme, the weighted sum of the execution delay and the task discarding cost is used as the execution cost of the calculation task, and the calculation task execution cost model is as follows:
wherein ,
represents the execution cost of the computation task of the tth time slot, is>
Represents a penalty not performed after the acquisition of the calculation task, based on the status of the evaluation unit>
A calculation delay representing that the marine sensor device performs a calculation task locally within the tth time slot, and->
A transmission delay representing the offloading of the computing task to the fog node by the marine sensing device in the tth time slot, and->
=1 and />
=1 indicates that at the t-th time slot, respectively, the requested computation task is executed on the marine sensor device and the computation task is offloaded to the fog node.
According to a further technical scheme, in each time slot, part of energy reaching the ocean sensing equipment based on the energy collection technology is recorded as
And constructing an energy collection model, which is as follows:
wherein ,
representing the energy arriving at the marine sensing device at the beginning of the t-th time slot.
According to the further technical scheme, the energy consumption model of the ocean sensing equipment is energy consumed by the ocean sensing equipment in the t-th time slot, and the formula is as follows:
and satisfies the energy relationship constraint:
wherein ,
represents the battery energy level of the marine sensor device at the tth time slot, ->
Flags representing calculated unload models>
=1 and />
=1 indicates that at the t-th time slot, the requested computation task is performed on the marine sensor device and the computation task is offloaded to the fog node, and/or>
Represents a predetermined frequency of the CPU cycle at the current tth time slot, based on the comparison of the measured value of the time slot>
Represents the transmitting power of the ocean sensing equipment accessing the fog node, and>
indicates that the energy consumption of the marine sensor device to locally perform a calculation task in the tth time slot, is greater than or equal to>
Representing the energy consumption of the marine sensing device to offload the computational task to the fog node during the t-th time slot.
In a further technical scheme, the calculation unloading model is as follows:
s.t.
wherein ,
represents the maximum transmission power in the tth time slot, is greater than>
Represents a predetermined frequency of the maximum CPU cycle at the tth time slot, based on the comparison of the measured value of the CPU frequency value and the value of the timer value>
Represents the computation task execution cost for the T-th slot, T =0,1,. T-1,T represents the total number of slots,. And >>
Represents the battery energy level of the marine sensing device at the current tth time slot, and->
Flags representing calculated unload models>
=1 and />
=1 indicates that at the t-th time slot, the requested computation task is performed on the marine sensor device and the computation task is offloaded to the fog node, and/or>
Represents a predetermined frequency of the CPU cycle at the current tth time slot, based on the comparison of the measured value of the time slot>
Indicating the transmission power in conjunction with the access of the marine sensor device to the fog node>
Represents a fraction of energy reaching the marine sensing device, based on the measured energy level>
And the preset frequency of W CPU cycles required by the ocean sensing equipment to complete the calculation task at the t time slot is shown.
In a further technical scheme, the method for solving the optimal calculation unloading decision in the current time slot by using a dynamic algorithm comprises the following steps:
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;
and solving the simplified calculation unloading model to obtain an optimal calculation unloading decision in the current time slot.
According to a further technical scheme, the solving comprises the following steps:
when the t-th time slot begins, acquiring a task request indicated value of a calculation task requested by the current time slot marine sensing equipment
Battery energy level of a marine sensing device->
And the channel power of the ocean sensing equipment accessing the fog node->
；
Substituting the obtained information of the current time slot into the simplified calculation unloading model for solving, and determining the index of the calculation unloading model of the current time slot
Predetermined frequency of CPU cycles>
The transmitting power of the ocean sensing equipment accessing the fog node is->
Partial energy based on energy harvesting techniques to a marine sensing device>
；
And updating a battery energy queue of the marine sensing equipment according to the calculation determination result, judging whether the battery energy queue reaches a stable value of the current time slot, if so, outputting the calculation determination result as an optimal calculation unloading decision in the current time slot, and otherwise, calculating the optimal calculation unloading decision of the next time slot.
In a second aspect, the present disclosure provides a marine sensor network computing offloading system based on energy harvesting techniques.
An energy harvesting technology-based marine sensor network computing offloading system, comprising:
the information acquisition module is used for acquiring real-time information of the ocean sensing equipment in the current time slot in each time slot, wherein the real-time information comprises a task request indicating value of a calculation task requested by the ocean sensing equipment, channel power of an access fog node of the ocean sensing equipment and battery energy level of the ocean sensing equipment;
the model building module is used for building the computation delay and the energy consumption of the ocean sensing equipment for executing the computation task in the current time slot, unloading the computation task to the transmission delay and the energy consumption of the fog node, building a computation task execution cost model, an ocean sensing equipment energy consumption model and an energy collection model, and building a computation unloading model by taking the minimum long-term average execution cost as a target function and combining the built model;
and the unloading decision solving module is used for solving the calculation unloading model by using a dynamic algorithm according to the acquired real-time information, and solving to obtain the optimal calculation unloading decision of the current time slot.
The above one or more technical solutions have the following beneficial effects:
1. the invention provides a method and a system for calculating and unloading an ocean sensing network based on an energy collection technology, which are characterized in that real-time information of the ocean sensing equipment in each time slot is obtained, calculation delay and energy consumption of the ocean sensing equipment for executing a calculation task in the time slot and transmission delay and energy consumption for unloading the calculation task to a fog node are established, a calculation task execution cost model, an ocean sensing equipment energy consumption model and an energy collection model are established, the established models are integrated, a calculation unloading model is established by taking the minimum long-term average execution cost as an objective function, an optimal calculation unloading decision in the current time slot is solved by using a dynamic algorithm, an optimal calculation unloading strategy is provided for the ocean sensing network applying the energy collection technology, a proper task unloading proportion is determined for the ocean sensing equipment and the fog node, the calculation performance of the ocean sensing network is improved, and node resource waste is avoided.
2. The invention considers the problem that the ocean sensing equipment has limited resources and can not meet the memory requirement of the optimal strategy, provides a dynamic algorithm for solving, simplifies a calculation unloading model by utilizing a Lyapunov drift function and a Lyapunov drift plus penalty function, carries out iterative solution according to the obtained real-time information, and obtains the current optimal calculation unloading decision, thereby carrying out resource allocation, optimizing the resource allocation result, reducing the long-term average execution cost and improving the user service quality of the system.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, are included to provide a further understanding of the invention, and are incorporated in and constitute a part of this specification, illustrate exemplary embodiments of the invention and together with the description serve to explain the invention and not to limit the invention.
FIG. 1 is a flowchart of a method for calculating and unloading a marine sensor network based on an energy collection technology according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a computational offload model in an embodiment of the present invention;
FIG. 3 is a schematic diagram of energy harvesting in an embodiment of the present invention.
Detailed Description
It is to be understood that the following detailed description is exemplary and is intended to provide further explanation of the invention as claimed. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of exemplary embodiments according to the invention. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, and it should be understood that when the terms "comprises" and/or "comprising" are used in this specification, they specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof, unless the context clearly indicates otherwise.
Example one
The embodiment provides a marine sensor network computing unloading method based on an energy collection technology, as shown in fig. 1, including:
step S1, in each time slot, acquiring real-time information of ocean sensing equipment in the current time slot, wherein the real-time information comprises a task request indicated value of a calculation task requested by the ocean sensing equipment, channel power of an access fog node of the ocean sensing equipment and battery energy level of the ocean sensing equipment;
s2, constructing the calculation delay and energy consumption of the ocean sensing equipment for executing the calculation task in the current time slot, and constructing the transmission delay and energy consumption for unloading the calculation task to the fog node;
s3, constructing a calculation task execution cost model, an ocean sensing equipment energy consumption model and an energy collection model, and constructing a calculation unloading model by taking the minimum long-term average execution cost as a target function and combining the constructed models;
and S4, solving a calculation unloading model by using a dynamic algorithm according to the acquired real-time information, and solving to acquire an optimal calculation unloading decision of the current time slot, wherein the calculation unloading decision is used for representing an execution mode of a calculation task requested by the marine sensing equipment.
The embodiment considers a computation offloading model composed of a plurality of ocean sensing devices with computation power and fog nodes, as shown in fig. 2, the fog nodes are located d meters away from the ocean sensing devices and are in communication connection with an ocean observation network, and the ocean sensing devices (such as the ocean sensing devices A, B, C shown in fig. 2) can be ships, coastal base stations and the like, and can access the fog nodes through wireless channels. The marine sensing equipment is associated with the node server and runs a virtual machine, computing tasks representing the sensing equipment are executed, and partial computing tasks are transferred to the node to be executed, so that the QoS (quality of service) can be remarkably improved.
In the step S1, the real-time information of the marine sensor device in the current time slot is obtained. The real-time information comprises a task request indicated value of a calculation task requested by the ocean sensing equipment
The channel power of the ocean sensing equipment accessing the fog node is->
And a battery energy level of the marine sensing device->
. According to the acquired information, the constructed calculation unloading model is solved, and the optimal calculation unloading decision in each time slot is obtained through solving。
In this embodiment, the time is sliced into a plurality of time slots (or time slices), each time slot having a length of
Denoted as Γ e {0,1. }, the wireless channel remains static within each time slot, but varies between different time slots. With A (L,; based on the total weight of the cells)>
) Represents a calculation task, wherein L (in bits, bit) represents the data size of the calculation task and->
Indicating an expiration time for the completion of the computing task, i.e., computing task A (L,/>)>
) Need to be timed>
And (4) completing the process. An application program is run on the marine sensing device, and the computing task requested by the program is modeled as a Bernoulli process. At the beginning of each time slot, compute task a (L,
) A probability of ρ is requested and, correspondingly, a probability of (1- ρ) is not requested. If a compute task is requested during the tth time slot, the task request indicates a value ≥ er>
=1, otherwise->
=0, i.e. P (/ H)>
=1)=1-P(/>
=0)=ρ，t∈Γ，P(/>
= 1) = ρ denotes the probability ρ, P (£ P @) when a computing task is requested>
= 0) =1- ρ indicates that the probability when a calculation task is not requested is 1- ρ. In this embodiment, no buffer is available for calculating the request queue, and the application running on the marine sensor device is a delay-sensitive application whose execution time of the task of interest is not greater than the time slot length, i.e.,' greater than or equal to>
。/>
Acquiring channel power of ocean sensing equipment accessing fog nodes at the tth time slot
, wherein ,/>
∽/>
(x) And t ∈ Γ, <' >>
(x) Is->
A Cumulative Distribution Function (CDF).
After the real-time information of the marine sensing equipment is obtained, step S2 is executed, and according to the obtained real-time information of the marine sensing equipment, the calculation delay and the energy consumption of the marine sensing equipment for executing the calculation task in the time slot are constructed, and the transmission delay and the energy consumption of the marine sensing equipment for unloading the calculation task to the fog node in the time slot are constructed.
Specifically, each computational task may be performed locally on the marine sensing device or off-loaded to the fog node for execution thereby. Order to
E {0,1}, j = { m, s }, will = { (m, s })>
As an indicator for calculating an unloading model, wherein>
=1 and />
=1 then respectively indicate that at the t-th time slot, the requested computational task is performed on the marine sensor device or offloaded to the fog node, specifically when
If =1, then>
=0, indicating that at the t-th time slot, the requested computing task is performed on the marine sensing device; when/is>
If =1, then>
And =0, indicating that the requested computing task is offloaded to the fog node for execution at the t-th time slot. Therefore, the computational offload model should satisfy:
the marine sensing device requires a predetermined frequency of multiple CPU cycles to accomplish this computational task. The number of CPU cycles required to process 1bit size input data is denoted X, which can be obtained from different applications using off-line measurement methods, and the calculation task a (L,
) Then W = LX CPU cycles are required. W CPU cycles are needed to complete the computation task in the t-th time slotRecording the predetermined frequency of W CPU cycles in the tth time slot as ^ greater than or equal to>
I.e. one slot corresponds to a predetermined frequency, which can be achieved by adjusting the chip voltage by DVFS techniques.
Based on the preset frequency of W CPU cycles required by the marine sensing equipment to complete the calculation task at the t time slot
Constructing a calculation delay for the marine sensing device to perform a calculation task at time t ≧>
The method comprises the following steps:
accordingly, energy consumption of constructing a marine sensing device to perform a computational task locally at time t
The method comprises the following steps:
where k is the effective switched capacitor, and its value depends on the chip architecture.
To reduce the amount of computation when moving is performed, compute task a (L,
) And unloading to the fog node. In the present embodiment, it is assumed that the amount of input data of the calculation task is small, and therefore the transmission delay of the feedback can be ignored. Channel power &basedon ocean sensing device access fog node>
And an emission power->
Constructing transmission delay for unloading the calculation task to the fog node by the marine sensing equipment in the tth time slot according to the Shannon formula>
The method comprises the following steps:
wherein, L represents the input data amount of the calculation task, r represents the transmission rate of the t-th time slot, and:
in the above equation, ω represents a channel bandwidth and σ represents a gaussian noise power inside the channel.
Correspondingly, energy consumption of the ocean sensing equipment for unloading the calculation task to the fog node in the t-th time slot is constructed
The method comprises the following steps:
in the step S3, on the basis of obtaining the time delay and the energy consumption of the marine sensing device for executing the calculation task under different conditions through the step S2, a calculation task execution cost model, a marine sensing device energy consumption model and an energy collection model are first constructed.
In devices applying energy harvesting techniques, the design of the computational offloading strategy is more complex than traditional mobile cloud computing systems, with battery energy levels being time dependent, such that system decisions are coupled at different time slots. Therefore, an optimal computation offload policy should strike a good balance between the computational performance of the current and future computational tasks.
The task execution delay is one of the key indicators of the user QoS (Quality of service), and the present embodiment uses QoS to calculate the offloading policy. Due to the intermittency and sporadic nature of the collected energy, some of the requested computing tasks may not be performed and have to be abandoned, for example due to insufficient local computing energy, while the wireless channel from the marine sensing device to the fog node is in a deep fade, i.e. the task cannot be offloaded. In this regard, each abandoned computing task is penalized in cost units. Therefore, the present embodiment constructs a computation task execution cost model, and defines the execution cost as a weighted sum of the execution delay and the task discard cost, which can be expressed by the following formula:
wherein ,
the penalty which is not executed after the calculation task is acquired is represented, and actually, the penalty also belongs to the task execution cost.
As shown in fig. 3, the energy collection process can be regarded as a continuous energy packet reaching the ocean sensing device, specifically, wind energy, tidal energy, and solar energy are converted into electric energy by the transducer and the conversion circuit, and the electric energy is directly or indirectly stored in the energy storage device (such as the ocean sensing device described in this embodiment) by the power management circuit. Energy to the ocean sensing device at the beginning of the tth time slot by energy harvesting
In which different time slots &>
Is maximum of->
. In each time slot, the fraction of energy reaching the marine sensor device based on the energy harvesting technique is recorded as ^ er>
And constructing an energy collection model, which is as follows:
wherein ,
，/>
the maximum of the energy reached in all time slots.
The energy is collected and stored in the battery of the marine sensing device and can be used for local calculations or calculation offloading in the next time slot.
Furthermore, an energy consumption model of the marine sensing device is constructed, that is, the energy consumed by the marine sensing device in the t-th time slot is:
while satisfying the following energy relationship constraints:
wherein ,Brepresents the battery energy level of the ocean sensing equipment,
indicating the battery energy level of the marine sensing device at the t-th time slot (i.e., the current time slot).
Thus, the energy level change of the marine sensing device battery can be described as:
and then, constructing a calculation unloading model by taking the minimum long-term average execution cost as an objective function and combining the constructed model.
In particular, the problem will be optimized
The calculation unloading model is constructed by taking the minimum long-term average execution cost as an objective function, and is as follows:
s.t.
in the above formula, the first and second carbon atoms are,
represents the maximum transmission power (i.e., maximum transmit power) at the tth time slot, and->
Denotes the maximum CPU scheduled frequency at the T-th slot, T denotes the total number of slots, T =0,1.
And after the calculation unloading model is constructed, executing a step S4, namely solving the calculation unloading model by using a dynamic algorithm according to the acquired real-time information, and solving to obtain the optimal calculation unloading decision of the current time slot.
In the considered offloading system, the system state consists of task requests, collectable energy, battery energy level and channel state, the actions are energy collection and calculating offloading decisions, including CPU cycle scheduling frequency and allocated transmit power. The operation depends only on the current system state. Further, since the objective function is to minimize the long-term average execution cost, there is a problem in that
In effect, is a Markov Decision Process (MDP) problem. In principle, the question->
The optimal solution can be obtained by a standard MDP algorithm, such as a relative value iterative algorithm and a linear programming reformulation method. However, both algorithms require the use of finite states to describe the system and discretize the set of feasible actions. Since the MDP algorithm is based on numerical iterations, it is difficult to obtain a solution. In addition, the quantized state and the feasible action set may cause severe performance degradation, and the marine sensing device resources are very limited, and cannot meet the memory requirement for storing the optimal strategy. Thus, the present embodiment proposes a dynamic solution algorithm to solve the problem @>
。
A computation offload decision (i.e., a solution of a computation offload model) can be described by a multidimensional vector at a certain time node, and a Lyapunov function (Lyapunov function) is a non-negative and scalar expression of a state represented by the vector at most. If the system is developed towards an unexpected direction, the Lyapunov function becomes larger, so that the Lyapunov function approaches to 0 along the negative direction of the x axis to be stable, and when the Lyapunov function is stable, the calculation unloading decision (namely, a solution formed by multi-dimensional vectors) at the moment is considered as the optimal decision of the current time node.
In this embodiment, first, based on the battery energy level of the marine sensing device at the current time slot, a lyapunov function is constructed, as shown in the following formula:
then, a lyapunov drift function is constructed, which is specifically shown as the following formula:
introducing a expectation in the Lyapunov drift function, the expectation being expressed as:
in the above formula, the first and second carbon atoms are,
indicating that it is desired.
Naturally, it is also possible to use,
where C is a bounded constant greater than 0.
Due to solving the problem
(i.e., calculating the objective function of the offload model) an optimal solution is obtained that is computationally complex and difficult to calculate, and thus appropriately amplified in the manner described above to form questions +>
By solving the problem->
Get the question->
The optimal solution set of (2). The question formed is>
Comprises the following steps:
s.t.
in the problem of
The method is simplified again on the basis of (1), namely, the Lyapunov drift functions of all time slots are accumulated to obtain:
in the above formula, the first and second carbon atoms are,
represents the battery energy level of the marine sensor device at time slot 0, i.e., the initial battery energy level of the marine sensor device, typically { [ MEANS ])>
=0；/>
And the battery energy level of the ocean sensing equipment at the Tth time slot after T time slots (namely 0~T-1 which are T time slots) are passed is shown.
Taking the expectation for both sides of the above equation, we get:
to make the above formula hold, let
Then the question can be picked up>
The method comprises the following steps:
s.t.
thus, the lyapunov drift function and the lyapunov drift plus penalty function can be expressed as:
the minimum value is determined in each time slot by means of weightsVTo adjust the pair
and />
Thus creating a new question->
Namely:
s.t.
by the scheme, the calculation unloading model is simplified by utilizing the Lyapunov drift function and the Lyapunov drift plus penalty function, and the problem that the optimal solution is difficult to solve
Switch value question->
By solving the problem->
And acquiring a solution set which is almost the same as the optimal solution and contains the optimal solution.
After the simplification of the computational offload model is completed, the problem is solved by the following algorithm
The optimal calculation unloading decision in the current time slot is solved.
Firstly, when the t-th time slot starts, a task request indicating value of a calculation task requested by current time slot marine sensing equipment is obtained
(task request indicator value->
The value of (5) is {0,1} which represents whether a calculation task is acquired or not), and the battery energy level of the ocean sensing equipment->
And the channel power of the ocean sensing equipment accessing the fog node->
。
Secondly, substituting the obtained information of the current time slot into the objective function of the simplified calculation unloading model for solving, solving to obtain a solution consisting of multidimensional vectors, namely determining the index of the calculation unloading model
Predetermined frequency of CPU cycles>
Transmitting power of an ocean sensing device accessing a fog node->
Partial energy based on energy harvesting techniques to a marine sensing device>
. Wherein, the simplified calculation unloading model is shown as the following formula:
s.t.
then, the battery energy queue of the ocean sensing equipment is updated according to the calculation determination result, namely the battery energy queue is updated
Judging whether the stable value of the current time slot is reached, namely judging whether the constraint condition of the calculation unloading model is met, if so, considering that the stable value of the current time slot is reached, and outputting the result determined by calculation as the optimal calculation unloading decision in the current time slot; otherwise, calculating the optimal calculation unloading decision of the next time slot.
In consideration of the limitation of resources (computation, storage and energy) of ocean sensing equipment and the dynamic topology of ocean nodes, the traditional unloading method cannot be applied, so that the embodiment provides a dynamic computation unloading strategy based on Lyapunov (Lyapunov) optimization, the strategy is low in complexity, the weak computation capability of the ocean nodes is adapted, the energy and the computation capability of each ocean node are fully utilized, the computation task is more reasonably distributed, the purpose of reducing average decision delay and system energy consumption is achieved, and the quality of service (QoS) is ensured.
Example two
The embodiment provides a marine sensor network computing unloading system based on an energy collection technology, which comprises:
the information acquisition module is used for acquiring real-time information of the ocean sensing equipment in the current time slot in each time slot, wherein the real-time information comprises a task request indicating value of a calculation task requested by the ocean sensing equipment, channel power of an access fog node of the ocean sensing equipment and battery energy level of the ocean sensing equipment;
the model building module is used for building the computation delay and the energy consumption of the ocean sensing equipment for executing the computation task in the current time slot, unloading the computation task to the transmission delay and the energy consumption of the fog node, building a computation task execution cost model, an ocean sensing equipment energy consumption model and an energy collection model, and building a computation unloading model by taking the minimum long-term average execution cost as a target function and combining the built model;
and the unloading decision solving module is used for solving the calculation unloading model by using a dynamic algorithm according to the acquired real-time information, and solving to obtain the optimal calculation unloading decision of the current time slot.
The steps related to the second embodiment correspond to the first embodiment of the method, and the detailed description thereof can be found in the relevant description of the first embodiment.
Those skilled in the art will appreciate that the modules or steps of the present invention described above can be implemented using general purpose computer means, or alternatively, they can be implemented using program code that is executable by computing means, such that they are stored in memory means for execution by the computing means, or they are separately fabricated into individual integrated circuit modules, or multiple modules or steps of them are fabricated into a single integrated circuit module. The present invention is not limited to any specific combination of hardware and software.
The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Although the embodiments of the present invention have been described with reference to the accompanying drawings, it is not intended to limit the scope of the present invention, and it should be understood by those skilled in the art that various modifications and variations can be made without inventive efforts by those skilled in the art based on the technical solution of the present invention.
Claims (10)
1. A marine sensor network calculation unloading method based on an energy collection technology is characterized by comprising the following steps:
in each time slot, acquiring real-time information of the current time slot marine sensing equipment; the real-time information comprises a task request indicating value of a calculation task requested by the ocean sensing equipment, channel power of an access fog node of the ocean sensing equipment and battery energy level of the ocean sensing equipment;
constructing the calculation delay and energy consumption of the ocean sensing equipment for executing the calculation task in the current time slot, and the transmission delay and energy consumption for unloading the calculation task to the fog node;
building a calculation task execution cost model, an ocean sensing equipment energy consumption model and an energy collection model, and building a calculation unloading model by taking the minimum long-term average execution cost as a target function and combining the built model;
and solving the calculation unloading model by using a dynamic algorithm according to the acquired real-time information, and solving to acquire the optimal calculation unloading decision of the current time slot.
2. The energy harvesting technology-based ocean sensing network computation offloading method of claim 1, wherein the predetermined frequency is based on W CPU cycles required for the ocean sensing device to complete the computation task at the t-th time slot
And constructing the calculation delay of the marine sensing equipment for locally executing the calculation task in the t time slot
And energy consumption
The formula is as follows:
wherein ,kin order to effectively switch the capacitance on and off,windicates the number of CPU cycles, W =1, …, W.
3. The energy harvesting technology-based ocean sensing network computational offloading method of claim 1 wherein channel power based on ocean sensing device access mist node
And transmit power
Constructing the transmission time delay for the ocean sensing equipment to unload the calculation task to the fog node in the t-th time slot according to the Shannon formula
And energy consumption
The formula is as follows:
wherein L represents the input data amount of the calculation task, r represents the transmission rate of the t-th time slot, and
in the above equation, ω represents a channel bandwidth and σ represents a gaussian noise power inside the channel.
4. The ocean sensing network computation offloading method based on energy harvesting technology of claim 1, wherein the weighted sum of the execution delay and the task abandon cost is used as the computation task execution cost, and then the computation task execution cost model is:
wherein ,
represents the computation task execution cost of the t-th slot,
represents a penalty not performed after the computation task is captured,
represents the computation delay of the marine sensing device to locally perform the computation task in the t-th time slot,
representing the transmission time delay of the ocean sensing equipment for unloading the calculation task to the fog node in the t-th time slot,
=1 and 
=1 indicates that at the t-th time slot, respectively, the requested computation task is executed on the marine sensor device and the computation task is offloaded to the fog node.
5. The energy harvesting technology based ocean sensing network computation offloading method of claim 1 wherein in each time slot, the energy harvesting technology based ocean will arrivePartial energy of the sensing device is noted
And constructing an energy collection model, which comprises the following steps:
wherein ,
representing the energy arriving at the marine sensing device at the beginning of the t-th time slot.
6. The energy collection technology-based ocean sensing network computation unloading method according to claim 1, wherein the energy consumption model of the ocean sensing equipment is the energy consumed by the ocean sensing equipment in the t-th time slot, and the formula is as follows:
and satisfies the energy relationship constraint:
wherein ,
represents the battery energy level of the marine sensing device at the t-th time slot,
an index representing a model of the computational load-shedding,
=1 and 
=1 indicates that at the t-th time slot, respectively, the requested computation task is executed on the marine sensor device and the computation task is offloaded to the fog node,
indicating the predetermined frequency of CPU cycles at the current t-th slot,
representing the transmitted power of the ocean sensing device accessing the fog node,
representing the energy consumption of the marine sensing device to perform the computation task locally during the t-th time slot,
representing the energy consumption of the marine sensing device to offload the computational task to the fog node during the t-th time slot.
7. The ocean sensing network computation offloading method based on energy harvesting technology of claim 1, wherein the computation offloading model is:
s.t.
wherein ,
represents the maximum transmission power at the t-th slot,
a predetermined frequency representing a maximum CPU cycle at the t-th slot,
represents the computation task execution cost of the T-th slot, T =0,1,. The T-1,T represents the total number of slots,
the indication is taken as to what is desired,
represents the battery energy level of the marine sensing device at the current t-th time slot,
an index representing a model of the computational load-shedding,
=1 and 
=1 indicates that at the t-th time slot, respectively, the requested computation task is executed on the marine sensor device and the computation task is offloaded to the fog node,
indicating the predetermined frequency of CPU cycles at the current t-th slot,
representing the transmitted power of the ocean sensing device accessing the fog node,
representing a portion of the energy reaching the ocean sensing equipment,
and the preset frequency of W CPU cycles required by the ocean sensing equipment to complete the calculation task at the t-th time slot is shown.
8. The ocean sensor network computation offloading method based on energy harvesting technology of claim 1, wherein solving the optimal computation offloading decision in the current time slot using a dynamic algorithm comprises:
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;
and solving the simplified calculation unloading model to obtain an optimal calculation unloading decision in the current time slot.
9. The energy harvesting technology-based ocean sensing network computing offloading method of claim 8, wherein the solving comprises:
when the t-th time slot begins, acquiring a task request indicated value of a calculation task requested by the current time slot marine sensing equipment
Battery energy level of marine sensing device
And the channel power of the ocean sensing equipment for accessing the fog node
；
Substituting the obtained information of the current time slot into the simplified calculation unloading model for solving, and determining the index of the calculation unloading model of the current time slot
Predetermined frequency of CPU cycles
Transmitting power of ocean sensing equipment access fog node
Partial energy reaching ocean sensing equipment based on energy collection technology
；
And updating a battery energy queue of the marine sensing equipment according to the calculation determination result, judging whether the battery energy queue reaches a stable value of the current time slot, if so, outputting the calculation determination result as an optimal calculation unloading decision in the current time slot, and otherwise, calculating the optimal calculation unloading decision of the next time slot.
10. A marine sensor network computation 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 ocean sensing equipment in the current time slot in each time slot, wherein the real-time information comprises a task request indicating value of a calculation task requested by the ocean sensing equipment, channel power of an access fog node of the ocean sensing equipment and battery energy level of the ocean sensing equipment;
the model building module is used for building the computation delay and the energy consumption of the ocean sensing equipment for executing the computation task in the current time slot, unloading the computation task to the transmission delay and the energy consumption of the fog node, building a computation task execution cost model, an ocean sensing equipment energy consumption model and an energy collection model, and building a computation unloading model by taking the minimum long-term average execution cost as a target function and combining the built model;
and the unloading decision solving module is used for solving the calculation unloading model by using a dynamic algorithm according to the acquired real-time information, and solving to obtain the optimal calculation unloading decision of the current time slot.
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 true CN115843070A (en) | 2023-03-24 |
| CN115843070B 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 (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170013495A1 (en) * | 2015-07-10 | 2017-01-12 | 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 |
| CN109829332A (en) * | 2019-01-03 | 2019-05-31 | 武汉理工大学 | A kind of combined calculation discharging method and device based on energy collection technology |
| 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 |
| CN113114733A (en) * | 2021-03-24 | 2021-07-13 | 重庆邮电大学 | Distributed task unloading and computing resource management method based on energy collection |
| CN114385272A (en) * | 2022-03-24 | 2022-04-22 | 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 |
| CN114650568A (en) * | 2022-03-18 | 2022-06-21 | 重庆邮电大学 | Distributed unloading method based on energy collection in mobile Ad Hoc cloud |
| CN115696451A (en) * | 2022-10-10 | 2023-02-03 | 广州大学 | Optimization method for energy and task scheduling of edge computing system |
| CN115686669A (en) * | 2022-10-17 | 2023-02-03 | 中国矿业大学 | Mine Internet of things intelligent computing unloading method assisted by energy collection |
-
2023
- 2023-02-23 CN CN202310152246.XA patent/CN115843070B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170013495A1 (en) * | 2015-07-10 | 2017-01-12 | 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 |
| CN109829332A (en) * | 2019-01-03 | 2019-05-31 | 武汉理工大学 | A kind of combined calculation discharging method and device based on energy collection technology |
| 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 |
| CN113114733A (en) * | 2021-03-24 | 2021-07-13 | 重庆邮电大学 | Distributed task unloading and computing resource management method based on energy collection |
| CN114650568A (en) * | 2022-03-18 | 2022-06-21 | 重庆邮电大学 | Distributed unloading method based on energy collection in mobile Ad Hoc cloud |
| CN114385272A (en) * | 2022-03-24 | 2022-04-22 | 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 |
| CN115686669A (en) * | 2022-10-17 | 2023-02-03 | 中国矿业大学 | Mine Internet of things intelligent computing unloading method assisted by energy collection |
Non-Patent Citations (4)
| Title |
|---|
| XIN LI等: "Energy-Efficient Data Collection Scheme Based on Mobile Edge Computing in WSNs", 《IEEE》 * |
| 唐伦;肖娇;魏延南;赵国繁;陈前斌;: "基于云雾混合计算的车联网联合资源分配算法", 电子与信息学报, no. 08 * |
| 李霆;: "移动边缘计算系统中的卸载和计算联合优化", 电脑知识与技术, no. 25 * |
| 蒋欣秀等: "面向海洋节能边缘计算的任务卸载研究", 《计算机工程与科学》, pages 1563 - 1573 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115843070B (en) | 2023-06-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Tak et al. | Federated edge learning: Design issues and challenges | |
| Zhu et al. | BLOT: Bandit learning-based offloading of tasks in fog-enabled networks | |
| CN113452956B (en) | Intelligent distribution method and system for power transmission line inspection tasks | |
| CN114866593B (en) | Industrial network control architecture with cooperative end edges and resource joint allocation method | |
| Tutuncuoglu et al. | Communicating using an energy harvesting transmitter: Optimum policies under energy storage losses | |
| CN112383931A (en) | Method for optimizing cost and time delay in multi-user mobile edge computing system | |
| CN116599602B (en) | Low-energy-consumption long-distance sonar communication system for ocean monitoring | |
| CN115175217A (en) | Resource allocation and task unloading optimization method based on multiple intelligent agents | |
| CN115065992A (en) | Edge calculation-based cooperative sensing and resource allocation method | |
| CN114385272B (en) | Ocean task oriented online adaptive computing unloading method and system | |
| Mekala et al. | Computational intelligent sensor-rank consolidation approach for industrial internet of things (iiot) | |
| Piorno et al. | HOLLOWS: A power-aware task scheduler for energy harvesting sensor nodes | |
| CN117156492A (en) | Deep reinforcement learning-based dual-time-scale resource allocation method for joint service caching, communication and calculation | |
| CN115843070B (en) | Ocean sensing network calculation unloading method and system based on energy collection technology | |
| CN112312299A (en) | Service unloading method, device and system | |
| Gambiroža et al. | Big data challenges and trade-offs in energy efficient internet of things systems | |
| CN114979027B (en) | Offshore edge computing network task unloading and resource allocation method based on energy collection | |
| Nassar et al. | Heterogeneous data reduction in WSN: Application to Smart Grids | |
| CN116109058A (en) | Substation inspection management method and device based on deep reinforcement learning | |
| CN112600869B (en) | Calculation unloading distribution method and device based on TD3 algorithm | |
| CN114077482B (en) | Intelligent computing optimization method for industrial intelligent manufacturing edge | |
| Abdelrahman | A Markov-modulated diffusion model for energy harvesting sensor nodes | |
| Wang et al. | Offloading Strategies for Mobile Edge Computing Based on Multi-Attribute Preferences in Smart Grids | |
| Chai et al. | Self-aware power management for maintaining event detection probability of supercapacitor-powered cyber-physical systems | |
| Dolui et al. | Towards on-board learning for harvested energy prediction |
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 |