CN114125061A - Resource optimal allocation method in shared edge service platform - Google Patents
Resource optimal allocation method in shared edge service platform Download PDFInfo
- Publication number
- CN114125061A CN114125061A CN202111196933.9A CN202111196933A CN114125061A CN 114125061 A CN114125061 A CN 114125061A CN 202111196933 A CN202111196933 A CN 202111196933A CN 114125061 A CN114125061 A CN 114125061A
- Authority
- CN
- China
- Prior art keywords
- platform
- edge service
- esp
- market
- edge
- 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 14
- 238000013468 resource allocation Methods 0.000 claims abstract description 26
- 230000014509 gene expression Effects 0.000 claims description 3
- 239000000126 substance Substances 0.000 claims description 3
- 230000035945 sensitivity Effects 0.000 claims description 2
- 239000012717 electrostatic precipitator Substances 0.000 claims 3
- 230000011218 segmentation Effects 0.000 claims 2
- 238000009827 uniform distribution Methods 0.000 claims 1
- 230000008901 benefit Effects 0.000 abstract description 7
- 238000004364 calculation method Methods 0.000 abstract description 4
- 230000000694 effects Effects 0.000 abstract description 4
- 230000002860 competitive effect Effects 0.000 abstract 1
- 238000010586 diagram Methods 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000002146 bilateral effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 239000011541 reaction mixture Substances 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses an optimal resource allocation method in a shared edge service platform, belongs to the technical field of calculation, calculation or counting, and particularly provides a pricing method for maximizing the profit of the edge service platform under a competitive relationship, and a utility function is defined to describe the satisfaction degree of a user and an ESP; setting constraint conditions to divide the network market into four situations; pricing that the required quantity of ESP resources reaches balance under four scenes is obtained respectively; and obtaining the optimal resource allocation scheme of the platform under four scenes according to the balanced pricing. The invention solves the problem that the prior resource allocation scheme ignores the platform competition effect, ensures the maximum platform benefit and simultaneously improves the feasibility of the resource allocation scheme.
Description
Technical Field
The invention relates to the field of Internet of things and an edge computing technology, particularly discloses an optimal resource allocation method in a shared edge service platform, and belongs to the technical field of computing, calculating or counting.
Background
In recent years, edge calculation is continuously developed, and the edge calculation is widely applied to the fields of intelligent transportation, smart cities, smart homes and the like by the characteristics of low time delay, high bandwidth, high reliability, massive connection, heterogeneous convergence and the like. Meanwhile, a plurality of servers in the network edge are used as network resources to provide various data processing and computing services for the equipment. However, edge computing is still in the early development stage, and with more and more devices connected to the internet, the demands of edge mobile devices such as smart phones and personal computers, data streams, and edge computing services are greatly increased. Servers deployed at the edge of the network provide corresponding edge services to users, and these edge servers are necessary to be effectively allocated as resources in the network. However, the existing edge computing resource allocation technology has certain limitations, which are mainly reflected in the problem that the existing MEC resource management and task scheduling decisions focus on the proportion of computing tasks executed locally or unloaded to the edge, the problem of selecting the transmission power when the user communicates with the base station, and the problem of which edge server the computing tasks are unloaded to. Aiming at the problems, the invention aims to analyze the resource allocation problem of the network platform by adopting a shared economic model, convert the resource allocation problem into the non-cooperative game problem among the platforms and realize the purpose of optimal resource allocation by researching the optimal pricing of each platform.
Disclosure of Invention
The invention aims to provide an optimal resource allocation method in a shared edge service platform, establish a shared economic model considering competition between platforms, analyze pricing of maximum profit of the two platforms under the competition relationship through a non-cooperative game, finally provide an optimal resource allocation scheme of the platforms, solve the technical problems of unreasonable resource allocation and low platform profit of the traditional network, and realize reasonable pricing of the shared edge service platform and optimal allocation of edge computing resource supply.
The invention adopts the following technical scheme for realizing the aim of the invention:
consider two operators, each offering an edge service platform, where multiple service providers offer data computation and processing services for users in a network. Thus, the platform can be viewed as an intermediary providing shared edge services, with a competing relationship between the two platforms. For the user, they need to select a platform to access and then pay a fee to obtain the edge services offered by the provider in the platform. The choice of the platform by the user depends mainly on factors such as own preference and pricing of the platform to the service. For Edge Service Providers (ESPs), initially after choosing to belong to either platform, it cannot be changed any more. Upon completion of each transaction, the platform will allocate a portion of the revenue to the service provider.
The optimal resource allocation scheme in the shared edge service platform comprises steps S1 to S4.
S1: defining utility function to describe satisfaction of user and ESP
S1.1: assuming that each user has different preference for platform selection, using l to represent the preference degree of the user for the platform, wherein the value of l is uniformly distributed in the interval [0,1], and the lower the value of l is, the more users select the platform 0; a higher value of l represents more user selections for platform 1. The user population is a unified body composed of a plurality of continuous users, and the preference degrees l of the users are uniformly distributed in the interval [0,1] with the intensity lambda. The utility function of platform i (i ═ 0 or 1, representing platform 0 and platform 1, respectively) selected by the user is as shown in equations (1) and (2):
in the formula (1) and the formula (2),
selecting utility function values of a platform 0 and a platform 1 for a user; p is a radical of0And q is0Respectively represent the price and quality of service (QoS), p, of the edge service in platform 00Representing the congestion degree of the platform 0, wherein the congestion degree represents the utilization rate of the system; p is a radical of1And q is1Respectively representing the price and quality of service (QoS), p, of the edge service in the platform 11Representing the congestion degree of the platform 1, wherein the congestion degree represents the utilization rate of the system; beta represents the sensitivity of the user to the degree of congestion of the system.
S1.2: the utility function of the ESPs belonging to platform i is:
in the formula (3), the reaction mixture is,
representing utility function values for ESPs in platform i,
representing the operating cost spent by the ESPs in platform i in striving for each transaction, and c representing the commission paid to the platform by each ESP. The net revenue from the ESP is a portion of the cost per transaction paid by the customer to the platform, denoted as α piWherein α ∈ [0,1]]Alpha is the proportion of the amount of each transaction obtained by the ESP, piThe price for edge service in platform i, from which the net revenue per transaction ESP is derived as api-c。
S2: setting constraint conditions to divide network market into four scenarios
S2.1: consider two schemes for platform-split markets:
in the first scenario, we assume that the user can only choose to access one platform or have another choice (i.e., neither platform is selected). If the user has other choices, it will produce a utility function with a value of zero, which can be derived from equation (1):
lwhich represents the maximum value of the user's preference for platform 0, the preference level is 0,l]selects platform 0. Similarly, from formula (2):
lrepresents the minimum value of the preference degree of the user to the platform 1, and the preference degree is
The user of (1) is selecting the platform with a preference degree of
Neither platform is selected. At this time, the two platforms do not occupy all the user resources of the market, and we define this Case as partially covering the market, referred to as Case P for short.
In the second scheme, the first step is that,
each user selects at least one platform, with preference degree
The user of (2) will select one of the two platforms, so the two platforms compete with each other for the portion of the user. Order to
The following can be obtained:
a threshold value representing the degree of user preference for selecting either platform 0 or platform 1. At this time, the two platforms occupy all the user resources of the market, and we define this situation as completely covering the market, called Case F for short.
Meanwhile, the demand of the edge service is defined as the number of users (D) who select the platform i to obtain the edge servicei(pi,pj) ) edgeThe supply amount of the edge service resource is the number (S) of ESPs that can provide the edge service in the platform ii(pi,pj))，pi，pjThe price for edge service in platform i, j (i, j ∈ {0,1}, i ≠ j),
in the formula (4), di(pi) Indicates the number of users selecting platform i, di(pi,pj) Indicating the number of users who finally select platform i in the situation where platform i and platform j compete with each other.
In addition, the pricing must satisfy the constraint Di(pi,pj)≤Si(pi,pj) The significance of this constraint is that the platform will try to increase its own edge service supply under any circumstances, so as to maintain the user's acceptance and satisfaction with the service, i.e. to ensure that there are sufficient resources to meet the user's needs. On this basis, the relationship between the supply amount and the demand amount can be divided into two cases:
the first case: supply is unconstrained, at which time Di(pi,pj)<Si(pi,pj) That is, the edge service supply of the platform is strictly greater than the demand of the edge service;
the second case: supply constraint, at this time Di(pi,pj)＝Si(pi,pj) I.e. the edge service supply of the platform is exactly equal to the demand of the edge service.
Therefore, under the interaction of edge service supply and demand, expressions when prices reach equilibrium in four parameter areas are obtained respectively, and optimal resource allocation schemes under different conditions are obtained according to the equilibrium prices, so that the regulation effect of exploration and utilization is achieved. The above constraints, combined with Case P and Case F, can divide the entire market into four scenarios:
wherein λ represents the intensity of the user preference parameter, which represents the market range, and the rectangular area formed by λ and l is the number of users selecting a certain platform in a coordinate system with l as the abscissa and λ as the ordinate.
S3: pricing for balancing ESP resource demand under four scenes
S3.1 platform gains can be expressed as:
Πi(pi,pj)＝[(1-α)pi+c]Di(pi,pj) (6)，
and i, j is belonged to {0,1}, i is not equal to j
The pricing when platform i revenue reaches maximum is as follows:
wherein the content of the first and second substances,
without some agreement being reached between the two platforms, the pricing strategy of the platform depends not only on itself but also on the other, i.e. non-cooperative gaming. At the moment, each platform only concerns whether the income of the platform reaches the maximum or not, and the equilibrium state reached under the condition of the non-cooperative game is called Nash equilibrium. Equation (7) is the pricing for platform i in four cases under equilibrium.
S4: obtaining optimal resource allocation scheme of platform under four scenes according to balanced pricing
S4.1 ESP resource supply Si *Pricing p with platform ii *The relationship of (a) to (b) is as follows:
in particular, p under (D) ofi＝pjFrom this, it follows that the ESP resource supply amount of the platform under the four pricing proposed at S3 is:
wherein the content of the first and second substances,
equation (9) is the resource allocation amount of the platform i under the balanced state in four situations, at this time, the respective profit of the platform reaches the maximum, and the profit of the platform is reduced when any party changes the pricing strategy.
By adopting the technical scheme, the invention has the following beneficial effects: the invention provides a resource allocation scheme for maximizing platform benefits, which considers the influence of resource supply and demand bilateral and the competition effect among platforms, optimizes the resource allocation scheme of the edge service by using a non-cooperative game based on a shared economic model, ensures the maximum platform benefits, improves the feasibility of the resource allocation scheme, and provides possibility for more intensive terminal devices to acquire the edge service.
Drawings
Fig. 1 is a network model of a shared edge service platform based on a shared economic model according to an embodiment of the present invention.
FIG. 2 is a diagram of a demand distribution diagram when two platforms independently exist in the market without affecting each other according to an embodiment.
FIG. 3 is a graph illustrating the demand distribution when two platforms compete with each other according to an embodiment.
FIG. 4 is a distribution diagram of the number of ESPs that can provide edge services in the platform according to an embodiment.
FIG. 5 is a diagram of four parameter regions that the price balance falls into under different condition settings provided by the first embodiment.
Detailed Description
In the following, the technical solutions in the embodiments of the present invention are clearly and completely described with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, fall within the scope of the present invention.
The embodiment provides an optimal resource allocation scheme in a shared edge service platform.
As shown in fig. 1, the network model based on the shared edge service platform under the shared economic model provided in this embodiment includes a platform 0 and a platform 1 of competing users, and after selecting a platform, the ESP provides an edge service for the user selecting the platform. The optimal resource allocation scheme in the shared edge service platform comprises 4 parts: utility functions of the user and the ESP; four scenarios for the network market; the required quantity of ESP resources under four scenes reaches balanced pricing; and (4) optimal resource allocation scheme of the platform under four scenarios.
Firstly, each user is defined to have a certain preference for any platform, denoted by l, the parameter is uniformly distributed in [0,1], the closer the value to 0, the more the user tends to select the platform 0, the closer to 1, the more the platform 1 is preferred, the parameter is the property of the user, and the platform can not be determined. The benefit function of the user can be derived by combining other factors influencing the user, such as price and congestion degree. For an ESP in a platform, its utility function is affected by its own revenue, mainly including service pricing, system congestion level, and operating costs. From this, the utility function of the ESP can be derived.
Next, the competing effects between the platforms are considered. When competition does not exist, the user selects either one of the platforms or neither of the platforms, the two platforms independently occupy one part of the market, and pricing decisions of each other do not influence the other side; when the platforms compete with each other, the users at leastOne of the two platforms is selected and pricing for the service affects the benefit of both the party and the other. As the implementation scenarios are shown in fig. 2 and fig. 3, fig. 2 and fig. 3 visually show the distribution of the user demand in the market. At the same time, the platform's own net revenue is also related to the demand for service. To prevent the loss of users and to maintain the normal operation of the platform, it is necessary to provide that the supply volume in the market is not less than the demand volume, and in particular, can be divided into supply constraints (D)i(pi,pj)＝Si(pi,pj) And supply unconstrained (D)i(pi,pj)<Si(pi,pj) Two cases). The market can be divided into four scenarios by combining the two constraints.
Then, in the above-mentioned four scenarios, in order to facilitate the research and analysis of the balanced price, the values of other irrelevant parameters are reasonably set: the platform distributes the amount paid by the user after each service transaction is completed to the ESP (namely alpha is 1); specifying QoS agreement for two platforms (i.e. q)i＝qjQ). An expression of the equilibrium price under four parameter settings is obtained through a static analysis method, a corresponding graph is obtained through MATLAB software simulation, and four parameter regions in which the equilibrium price falls under different condition settings are shown in FIG. 5.
And finally, under the condition that the equilibrium price is known, obtaining an optimal allocation scheme of the service resource supply quantity of the platform under four scenes, wherein the price at the moment is the price when the platform demand quantity reaches equilibrium, so that the scheme can ensure that the benefit of the platform is the maximum. The distribution of the number of ESPs that can provide edge services in the platform is shown in fig. 4.
The above embodiments are further described in detail for the purpose of illustrating the invention, the technical solutions and the advantages, it should be understood that the above embodiments are only illustrative and are not intended to limit the scope of the invention, and any modification, equivalent replacement or modification made within the spirit and principle of the invention should be included in the scope of the invention.
Claims (5)
1. The optimal resource allocation method in the shared edge service platform is characterized in that a utility function is adopted to describe the satisfaction degree of users and ESPs, the market is divided into four situations by considering the segmentation condition of the edge service platform on the market and the constraint relation between the edge service supply quantity and the edge service demand quantity, pricing for balancing the ESP resource demand quantity under the four situations is calculated respectively, and the optimal resource allocation scheme of the edge service platform under the four situations is obtained according to the pricing for balancing the ESP resource demand quantity under the four situations.
2. The method of claim 1, wherein the utility function describing the satisfaction of the user is:
the utility function describing the satisfaction of an ESP is:
wherein the content of the first and second substances,
selecting utility function values of a platform 0 and a platform 1 for a user, wherein l is the preference degree of the user to the platform, and l is a uniform distribution interval [0,1] with the intensity lambda]Inner, p0And q is0Price and quality of service, ρ, respectively, for edge services in platform 00Degree of congestion, p, for platform 01And q is1Price and quality of service, ρ, respectively, for edge services in platform 11Is the congestion level of platform 1, beta is the sensitivity of the user to the congestion level of the system,
for the utility function values of the ESPs in platform i,
the operating cost spent for each transaction is strived for by the ESPs in platform i, c the commission paid to the platform for each ESP, and α is the proportion of the amount of each transaction the ESP obtains.
3. The method for optimally allocating resources in the shared edge service platform according to claim 2, wherein the specific method for dividing the market into four cases in consideration of the market division condition of the edge service platform and the constraint relationship between the edge service supply quantity and the edge service demand quantity is as follows: the market segmentation situation of the edge service platform comprises a partial coverage market situation and a full coverage market situation, the constraint relation between the edge service supply quantity and the edge service demand quantity comprises two situations that the edge service supply quantity is strictly greater than the edge service demand quantity and the edge service supply quantity is equal to the edge service demand quantity, the edge service demand quantity is obtained by calculating the number of users of the selection platform,
the expression that divides the market into four cases is:
Di(pi，pj) Number of users, p, to select platform i between platform i and platform ji、pjPrice of edge service in platform i and platform j, i, j belongs to {0,1}, i ≠ j, di(pi) Indicates the number of users selecting platform i, di(pi，pj) Representing the number of users who finally select the platform i under the situation that the platform i and the platform j compete with each other, Case P is a partial coverage market, Case F is a complete coverage market, configuration (A) is a partial coverage market in which the edge service supply is strictly greater than the edge service demand, configuration (B) is a partial coverage market in which the edge service supply is equal to the edge service demand, configuration (C) is a complete coverage market in which the edge service supply is strictly greater than the edge service demand, and configuration (D) is an edge serviceThe service supply amount is equal to the full coverage market of the edge service demand amount.
4. The method of claim 3, wherein the method for calculating the pricing balancing the ESP resource demand under four situations is to calculate the pricing when the platform gains of the four situations are the maximum, and the pricing p when the platform gains of the platform i are the maximumi *Is composed of
5. The optimal resource allocation method in the shared edge service platform according to claim 4, wherein the method for obtaining the optimal resource allocation scheme of the edge service platform under four situations according to the pricing that the ESP resource demand under four situations reaches equilibrium comprises the following steps: calculating the resource supply quantity of the edge service platform under four situations according to the mathematical relationship between the ESP resource supply quantity and the pricing that the ESP resource demand quantity reaches the equilibrium under four situations, and calculating the ESP resource supply quantity s of the platform i when the ESP resource demand quantity reaches the equilibrium pricing under four situationsi *Is composed of
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111196933.9A CN114125061B (en) | 2021-10-14 | 2021-10-14 | Resource optimal allocation method in shared edge service platform |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111196933.9A CN114125061B (en) | 2021-10-14 | 2021-10-14 | Resource optimal allocation method in shared edge service platform |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114125061A true CN114125061A (en) | 2022-03-01 |
| CN114125061B CN114125061B (en) | 2023-12-08 |
Family
ID=80376102
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111196933.9A Active CN114125061B (en) | 2021-10-14 | 2021-10-14 | Resource optimal allocation method in shared edge service platform |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114125061B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018215046A1 (en) * | 2017-05-22 | 2018-11-29 | Telefonaktiebolaget Lm Ericsson (Publ) | Edge cloud broker and method therein for allocating edge cloud resources |
| CN111131395A (en) * | 2019-11-29 | 2020-05-08 | 苏州浪潮智能科技有限公司 | FPGA cloud platform task resource allocation device, method and equipment |
| US20200296054A1 (en) * | 2019-03-11 | 2020-09-17 | Wipro Limited | Method and system for providing network resource sharing in multi-access edge computing (mec) network |
| CN113269461A (en) * | 2021-06-09 | 2021-08-17 | 北京理工大学 | Game-based edge computing resource management method |
-
2021
- 2021-10-14 CN CN202111196933.9A patent/CN114125061B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018215046A1 (en) * | 2017-05-22 | 2018-11-29 | Telefonaktiebolaget Lm Ericsson (Publ) | Edge cloud broker and method therein for allocating edge cloud resources |
| US20200296054A1 (en) * | 2019-03-11 | 2020-09-17 | Wipro Limited | Method and system for providing network resource sharing in multi-access edge computing (mec) network |
| CN111131395A (en) * | 2019-11-29 | 2020-05-08 | 苏州浪潮智能科技有限公司 | FPGA cloud platform task resource allocation device, method and equipment |
| CN113269461A (en) * | 2021-06-09 | 2021-08-17 | 北京理工大学 | Game-based edge computing resource management method |
Non-Patent Citations (1)
| Title |
|---|
| 王磊;***;康彬;崔景伍;郑宝玉;: "多主多从Stackelberg博弈下的边缘缓存资源分配算法", 信号处理, no. 04 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114125061B (en) | 2023-12-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Yi et al. | Two-stage spectrum sharing with combinatorial auction and stackelberg game in recall-based cognitive radio networks | |
| CN111757354A (en) | Multi-user slicing resource allocation method based on competitive game | |
| CN110888687B (en) | Mobile edge computing task unloading optimal contract design method based on contract design | |
| CN113037876B (en) | Cooperative game-based cloud downlink task edge node resource allocation method | |
| CN105007287A (en) | public cloud resource dynamic configuration method and system | |
| CN109831788B (en) | Network resource dynamic allocation method and system for edge access scene | |
| Koutsopoulos et al. | Auction mechanisms for network resource allocation | |
| CN106954234A (en) | User's connection and virtual resource allocation method in a kind of super-intensive heterogeneous network | |
| Siew et al. | Dynamic pricing for resource-quota sharing in multi-access edge computing | |
| Zhang et al. | A double auction mechanism for virtual resource allocation in SDN-based cellular network | |
| Chen et al. | Balancing income and user utility in spectrum allocation | |
| Tong et al. | Stackelberg game-based task offloading and pricing with computing capacity constraint in mobile edge computing | |
| CN108990067A (en) | A kind of energy efficiency controlling method applied to super-intensive heterogeneous network | |
| CN114928612B (en) | Excitation mechanism and resource allocation method for collaborative offloading in mobile edge computing | |
| CN113938394B (en) | Monitoring service bandwidth allocation method and device, electronic equipment and storage medium | |
| Chouayakh et al. | PAM: A fair and truthful mechanism for 5G dynamic spectrum allocation | |
| Radonjić et al. | Responsive pricing modeled with Stackelberg game for next-generation networks | |
| Mondal et al. | Dynamic price-enabled strategic energy management scheme in cloud-enabled smart grid | |
| CN114125061A (en) | Resource optimal allocation method in shared edge service platform | |
| CN110650047B (en) | Main and distribution network integrated dynamic resource distribution method and device | |
| CN111148252A (en) | LTE electric wireless private network resource allocation method based on auction | |
| CN109451542A (en) | A kind of presell system model and Layered Learning algorithm based on spectral overhead | |
| Kwon et al. | Royalties vs. upfront lump-sum fees in data communication environments | |
| CN115358765A (en) | Micro cloud resource pricing method in micro cloud alliance environment based on game theory | |
| Sterz et al. | Multi-stakeholder service placement via iterative bargaining with incomplete information |
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 |