CN117331667A - Non-precise mixed critical task energy-saving scheduling method based on group task partition - Google Patents

Abstract

The invention relates to a non-precise mixed critical task energy-saving scheduling method based on group task partitioning, which divides a non-precise mixed critical task set gamma into a low-critical-level task set gamma _LO And a high key hierarchy task set Γ _HI For gamma _LO According to the utilization rate in the low modeNon-ascending sort is performed on Γ _HI Tasks according to high mode utilizationNon-ascending sort is performed; will be gamma _HI And Γ _LO The tasks of the group task are combined into a group task according to a one-to-one correspondence, and the tasks which cannot form the group task are remained as an independent unit; for group tasks, a high-key-level task is distributed firstly by using an optimal adaptation method, then a low-key-level task is distributed, and for non-group tasks, the optimal adaptation method is utilizedTask allocation; computing processor P _l Energy consumption optimization speed S of (2) _l And normalizing the energy consumptionRecalculating normalized energy consumption of the entire task setAccording to the method, the non-precise mixing task is distributed through the optimal adaptation method based on the combined task, so that the energy consumption of the system can be effectively reduced.

Description

Non-precise mixed critical task energy-saving scheduling method based on group task partition

Technical Field

The invention relates to real-time scheduling of a real-time system, an embedded system and a hybrid critical system, in particular to a non-precise hybrid critical task energy-saving scheduling method based on group task partitioning.

Background

To address the challenges posed by the rapid development of computer technology to the design of embedded systems; a hybrid critical system that integrates different levels of applications into the same shared platform is its direction of development. Hybrid critical systems include a variety of different critical-level tasks. The high-critical-level tasks need to be securely authenticated and must be guaranteed to be completed within a deadline, or else can lead to serious consequences and even disasters. The low-criticality level task allows for only a slight impact on the user experience when an deadline is occasionally missed. Unmanned aerial vehicles, automotive autopilot systems, and aircraft control systems are the most common hybrid critical systems.

With the continuous progress of the ultra-large scale integrated circuit technology, the unit energy consumption of the processor is gradually increased. In order to improve the computational performance of the system, the evolution from single processor systems to multiprocessor systems has evolved. However, there is relatively little research on multiprocessor mixed critical task scheduling, and existing research has focused mainly on classical mixed critical task models, i.e., directly discarding low critical level tasks when the system is in high mode; this is not in practical agreement with industrial production. In addition, the existing low-energy-consumption scheduling method is mainly aimed at the traditional embedded real-time system, and has less application research on the hybrid key system.

Disclosure of Invention

The invention aims to overcome the defects of the prior art and provide an energy-saving scheduling method for non-precise mixing key tasks based on group task partitioning, which can effectively reduce the energy consumption of a system.

In order to achieve the above purpose, the technical scheme of the invention is that an energy-saving scheduling method for non-precise mixing key tasks based on group task partitioning comprises the following steps:

step 1, scheduling a non-precise mixed critical periodic task Γ on a multiprocessor system P using a dynamic priority strategy; dividing the non-exact mixed critical task set Γ into low-level critical task sets Γ _LO And a high key hierarchy task set Γ _HI For gamma _LO According to the utilization rate in the low modeNon-ascending sort is performed on Γ _HI Tasks according to high mode utilizationNon-ascending sort is performed;

step 2, Γ is set _HI And Γ _LO The tasks of the group task are combined into a group task according to a one-to-one correspondence, and the tasks which cannot form the group task are remained as an independent unit;

step 3, distributing high-key-level tasks for the group tasks by using an optimal adaptation method, then distributing low-key-level tasks, and distributing tasks for non-group tasks by using the optimal adaptation method;

step 4, calculating the processor P _l Energy consumption optimization speed S of (2) _l And normalizing the energy consumptionCalculating normalized energy consumption of the whole task set>Wherein P is _l Is a processor in the multiprocessor system P.

Preferably, the step 1 specifically includes:

in a multiprocessor system p= { P ₁ ,P ₂ ,…,P _M Scheduling non-precision mixed critical periodic tasks Γ= { τ using dynamic priority policy on } ₁ ,τ ₂ ,…,τ _n -a }; wherein M is the number of processors, and the value of M is a positive integer; each mixed critical period task τ _i From triples { T _i ,ξ _i ,C _i -composition; wherein i is more than or equal to 1 and less than or equal to n, n represents the number of tasks in the task set, i is an integer, and M is less than n; t (T) _i For mixing critical periodic tasks τ _i Is a period of (2); zeta type toy _i For mixing critical periodic tasks τ _i Is denoted as xi _i = { HI, LO }, when mixingCombining critical periodic tasks τ _i Key hierarchy ζ of (a) _i When HI, it is a high-critical-level task; when mixing critical period task tau _i Key hierarchy ζ of (a) _i When=lo, it is a low critical level task; c (C) _i For mixing critical periodic tasks τ _i Worst-case execution times in different modes; c (C) _i (HI) and C _i (LO) are respectively mixed critical period tasks τ _i Worst-case execution times in high mode and low mode; high pattern means that there is a high critical hierarchy task τ _i At processor speed S, execution times all exceed C _i (LO)/S; low mode refers to any task τ _i At processor speed S, execution time does not exceed C _i (LO)/S; if mixing critical period task tau _i For high key hierarchy task, then C _i (HI)≥C _i (LO); if mixing critical period task tau _i For low critical level tasks, then C _i (HI)≤C _i (LO); dividing the non-exact mixed critical task set Γ into low-level critical task sets Γ _LO And a high key hierarchy task set Γ _HI ；And->Respectively, imprecise mixing critical tasks tau _i Utilization in low mode and high mode; to gamma _LO According to the utilization in low mode +.>Non-ascending sort is performed on Γ _HI According to the task of high mode utilization +.>Non-ascending sort is performed.

Preferably, the step 2 specifically includes:

will be gamma _HI First high-critical-level task and Γ in (1) _LO In (a) and (b)The first low-key-level task is combined into a group task, and gamma is calculated _HI The second high-critical-level task and Γ in (1) _LO The second low-key-level task in (1) is combined into a group task, and Γ is calculated _HI Third high-critical-level task and Γ in (1) _LO The third low-key-level task in the list is combined into a group task until only tasks which do not form the group task remain, and the tasks which do not form the group task remain as a separate unit.

Preferably, the step 3 specifically includes:

processor P in multiprocessor system P _l Is the remaining utilization ratio U of (2) _l Is by having been allocated to the processor P _l Tasks and waits to be allocated to processor P _l Task τ _i Determined, U _l The calculation formula of (2) is as follows:

wherein U is _l Representing processor P _l Is used for the residual utilization rate of the (a);and->Respectively represent task tau _i And has been allocated to the processor P _l The sum of the utilization rates of the above tasks in the low mode and the high mode; zeta type toy _i =lo and ζ _i =hi represents task τ respectively _i The method comprises the steps of providing a low-key-level task and a high-key-level task; when processor P _l Is the remaining utilization ratio U of (2) _l Parameter x which is equal to or greater than 0 and is feasible to schedule _lo And x _up Satisfy the relation 0 < x _lo ≤x _up When less than or equal to 1, the data is distributed to the processor P _l Task set scheduling of (1) is feasible when the processor P _l Is the remaining utilization ratio U of (2) _l < 0 or the relation 0 < x _lo ≤x _up When 1 is not established, the data is distributed to the processor P _l Is not feasible for task set scheduling; for group tasks, first makeHigh key level task xi using optimum adaptation method _i =hi assigned to processor P _k Judging low key level task xi _j =lo assigned to processor P _k If the task is feasible, the task is distributed to other processors meeting the scheduling conditions by using an optimal adaptation method, and if the task is not feasible, the task is distributed to the processors meeting the scheduling conditions by using the optimal adaptation method;

processor P _l Parameter x of scheduling feasibility of (a) _up The calculation formula of (2) is as follows:

wherein,representing processor P _l The sum of the utilization of the task set phi (l) in the high mode; />Representing processor P _l The sum of the utilization of the low critical level tasks in the task set phi (l) in the low-high mode; />Representing processor P _l The sum of the utilization of the low critical level tasks in the task set phi (l) in the high mode;

processor P _l Schedule feasible parameter x of (2) _lo The calculation formula of (2) is as follows:

wherein,representing processor P _l The sum of the utilization of the high critical level tasks in the task set phi (l) in the low mode.

Preferably, the step 4 specifically includes:

processor P _l Energy consumption optimization speed S of (2) _l The calculation formula is as follows:

wherein,representing processor P _l An upper bound on the utilization of task set phi (l) in low mode; />Representing processor P _l An upper bound on the utilization of task set phi (l) in high mode; s is S _crit Representing processor P _l Key speed of (2); processor P _l The upper bound of the utilization of the task set phi (l) in low mode +.>The calculation formula of (2) is as follows:

wherein x is a deadline parameter, and the calculation formula is as follows:

processor P _l Upper bound of utilization of task set phi (l) in high modeThe calculation formula of (2) is as follows:

processor P _l Critical speed S of (2) _crit The calculation formula of (2) is as follows:

wherein P is _ind Representing frequency independent active power; m represents a dynamic power index, which is a system correlation constant, and is usually more than or equal to 2; c (C) _ef Representing the effective switched capacitance;

computing processor P _l Normalized energy consumption of (2)The formula of (2) is as follows:

calculating normalized energy consumption of the whole task setThe formula of (2) is as follows:

as can be seen from the above description of the present invention, compared with the prior art, the present invention has the following advantages:

(1) The method of the invention saves about 9.3% of energy consumption compared with other mixed key period scheduling methods;

(2) By reducing the energy consumption of the hybrid key system, the production cost of the product can be reduced, and the reliability of the product can be improved.

The invention is described in further detail below with reference to the accompanying drawings and examples, but the invention is not limited to the examples.

Drawings

FIG. 1 is a schematic flow chart of the method of the present invention.

FIG. 2 is a graph showing the results of the energy consumption test of the present invention.

Detailed Description

The following describes and discusses the technical solutions in the embodiments of the present invention in detail with reference to the drawings. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.

Referring to fig. 1, the energy-saving scheduling method for the non-precise mixed critical tasks based on the group task partition comprises the following steps.

Step 1, scheduling a non-precise mixed critical periodic task Γ on a multiprocessor system P using a dynamic priority strategy; dividing the non-exact mixed critical task set Γ into low-level critical task sets Γ _LO And a high key hierarchy task set Γ _HI For gamma _LO According to the utilization rate in the low modeNon-ascending sort is performed on Γ _HI Tasks according to high mode utilizationNon-ascending sort is performed.

In a multiprocessor system p= { P ₁ ,P ₂ ,…,P _M Scheduling non-precision mixed critical periodic tasks Γ= { τ using dynamic priority policy on } (M is the number of processors, its value is a positive integer) ₁ ,τ ₂ ,…,τ _n -n is the number of tasks, the value of which is a positive integer); each non-critical mixed critical periodic task τ _i From triples { T _i ,ξ _i ,C _i The composition (i is more than or equal to 1 and less than or equal to n, i is an integer, n is more than M), T _i For imprecisely mixing critical periodic tasks τ _i Is a period of (2); zeta type toy _i For imprecisely mixing critical periodic tasks τ _i Is denoted as xi _i = { HI, LO }, when not exactly mixing critical period task τ _i Key of (2)Level xi _i When HI, it is a high critical level task, when the critical period task τ is not exactly mixed _i Key hierarchy ζ of (a) _i When=lo, it is a low critical level task; c (C) _i For imprecisely mixing critical periodic tasks τ _i Worst-case execution time in different modes, wherein C _i (HI) and C _i (LO) represents the imprecise mixing critical periodic task τ respectively _i Worst-case execution times in high mode and low mode; high pattern means that there is a high critical hierarchy task τ _i At processor speed S, execution times all exceed C _i (LO)/S; low mode refers to any task τ _i At processor speed S, execution time does not exceed C _i (LO)/S; if mixing critical period task tau _i For high key hierarchy task, then C _i (HI)≥C _i (LO); if mixing critical period task tau _i For low critical level tasks, then C _i (HI)≤C _i (LO); the dynamic priority policy refers to earliest deadline priority policy scheduling, which determines the priority of a task through the deadline of the task, and the higher the deadline is, the lower the priority of the task is, the lower the deadline is, and the higher the priority of the task is; if the deadlines of the tasks are the same, comparing the arrival time of the tasks, wherein the higher the task priority with small arrival time is, the lower the task priority with large arrival time is; if the deadline and the arrival time of the tasks are the same, the higher the priority of the tasks with small subscripts, the lower the priority of the tasks with large subscripts; tasks with high priorities are scheduled first. Dividing the non-exact mixed critical periodic task set Γ into a low-critical-level task set Γ _LO And a high key hierarchy task set Γ _HI ；And->Respectively, the mixed critical period task tau _i Utilization in low mode and high mode; c (C) _i (LO) and C _i (HI) represents the hybrid critical cycle task τ respectively _i Execution time in low mode and high mode; t (T) _i For mixing critical periodic tasks τ _i Is a period of (2); for the low key hierarchy task set Γ _LO Use of utilization in low mode +.>Non-ascending sort is performed such that +.>(n ₁ For a low key hierarchy task set Γ _LO The number of tasks in) for a high key hierarchy task set Γ _HI Utilization in high mode>Non-ascending sort is performed such that +.>(n ₂ For a high key hierarchy task set Γ _HI The number of tasks in (a).

Step 2, Γ is set _HI And Γ _LO The tasks of the group task are combined into a group task according to a one-to-one correspondence, and the tasks which cannot form the group task are remained as an independent unit.

Will be gamma _HI First high-critical-level task and Γ in (1) _LO The first low-key-level task in (1) is combined into a group task, and Γ is calculated _HI The second high-critical-level task and Γ in (1) _LO The second low-key-level task in (1) is combined into a group task, and Γ is calculated _HI Third high-critical-level task and Γ in (1) _LO The third low-key-level task in the list is combined into a group task until only tasks which do not form the group task remain, and the tasks which do not form the group task remain as a separate unit.

And step 3, distributing high-key-level tasks for the group tasks by using an optimal adaptation method, then distributing low-key-level tasks, and distributing tasks for non-group tasks by using the optimal adaptation method.

The best adaptation method is always according to any oneThe order in which the utilization of tasks decreases allocates tasks and minimizes the remaining utilization of the processor; processor P _l Is the remaining utilization ratio U of (2) _l By having been allocated to the processor P _l Task on and waiting to be distributed to processor P _l Task tau on _i Determined, processor P _l Is the remaining utilization ratio U of (2) _l The calculation formula of (2) is as follows:

wherein,and->Respectively represent task tau _i And has been allocated to the processor P _l The sum of the utilization rates of the above tasks in the low mode and the high mode; zeta type toy _i =lo and ζ _i =hi represents task τ respectively _i The method comprises the steps of providing a low-key-level task and a high-key-level task; phi (l) represents task tau _i And has been allocated to the processor P _l Is a collection of tasks; when processor P _l Is the remaining utilization ratio U of (2) _l 0. Gtoreq.and processor P _l Scheduling feasibility parameter x _up And x _lo Satisfy 0 < x _lo ≤x _up When the relation is less than or equal to 1, the relation is distributed to the processor P _l The task set on the platform is feasible to dispatch; when processor P _l Is the remaining utilization ratio U of (2) _l < 0 or processor P _l Scheduling feasibility parameter x _up And x _lo Not satisfy 0 < x _lo ≤x _up When the relation is less than or equal to 1, the relation is distributed to the processor P _l Task set scheduling on the platform is not feasible; processor P _l Parameter x of scheduling feasibility of (a) _up The calculation formula of (2) is as follows:

wherein,representing processor P _l The sum of the utilization of the task set phi (l) in the high mode; />Representing processor P _l The sum of the utilization of the low critical level tasks in the task set phi (l) in the low mode; />Representing processor P _l The sum of the utilization of the low critical level tasks in the task set phi (l) in the high mode; />The calculation formula of (2) is as follows:

the calculation formula of (2) is as follows:

the calculation formula of (2) is as follows:

processor P _l Scheduling a feasible parameter x _lo The calculation formula of (2) is as follows:

wherein,and->Respectively represent the processors P _l The sum of the utilization rates of the high-key-level tasks and the low-key-level tasks in the task set phi (l) in the low mode; />The calculation formula of (2) is as follows:

for group tasks, the high-key-hierarchy task (τ) is first applied to the best-fit approach ₁ ∈Γ _HI ) Minimum remaining utilization processor P allocated to meeting scheduling conditions _k On the above, the task of the low key hierarchy (tau ₂ ∈Γ _LO ) Distributed to processor P _k If the task is not feasible, the task is distributed to other processors with minimum residual utilization rate meeting the scheduling conditions by using an optimal adaptation method, and for non-group tasks, the task is distributed to the processors with minimum residual utilization rate meeting the scheduling conditions by using the optimal adaptation method.

Step 4, calculating the processor P _l Energy consumption optimization speed S of (2) _l And normalizing the energy consumptionCalculating normalized energy consumption of the whole task set>

Processor P _l Energy consumption optimization speed S of (2) _l The calculation formula of (2) is as follows:

wherein,representing processor P _l An upper bound on the utilization of the task set phi (l) in the low mode; />Representing processor P _l The upper bound of the utilization of the task set phi (l) in the high mode; s is S _crit Representing processor P _l Key speed of (2); c (C) _i (LO) and C _i (HI) represents the mixed cycle task τ respectively _i Execution time in low mode and high mode; t (T) _i Representing mixed critical period task τ _i Is a period of (2); processor P _l The upper bound of the utilization of the task set phi (l) in low mode is +.>The calculation formula of (2) is as follows:

wherein x is a parameter of the deadline, and the calculation formula is as follows:

processor P _l Scheduling a feasible parameter x _up The calculation formula is as follows:

processor P _l Scheduling a feasible parameter x _lo The calculation formula is as follows:

processor P _l Upper bound of utilization of task set phi (l) in high modeThe calculation formula of (2) is as follows:

processor P _l Critical speed S of (2) _crit The calculation formula of (2) is as follows:

wherein P is _ind Representing frequency independent active power; m represents a dynamic power index (m.gtoreq.2); c (C) _ef Representing an effective capacitance switch; adopts the common setting P _ind ＝0.1，m＝3，C _ef =1, thereby obtaining S _crit ＝0.17。

According to processor P _l Energy consumption optimization speed S of (2) _l Calculation processor P _l Normalized energy consumption of (2)The formula of (2) is as follows:

wherein P is _ind Representing frequency independent active power; m represents a dynamic power index (m.gtoreq.2); c (C) _ef Representing an effective capacitance switch; adopts the common setting P _ind ＝0.1，m＝3，C _ef ＝1；C _i (LO) represents a mixed cycle task τ _i Execution time in low mode; t (T) _i Representing mixed critical period task τ _i Is a periodic one.

Based on normalized energy consumption of each processorCalculating normalized energy consumption of the whole task setThe formula of (2) is as follows:

where M represents the number of processors,representing normalized power consumption of each processor;

in this embodiment, the number of processors is set to 4, and the high-key-level tasks account for 50% of the total task set; the influence of the average utilization rate of each processor on the energy consumption in the low mode is examined; referring to FIG. 2, two methods were compared in an experiment, first, the method of the present invention; second, a critical hierarchy aware best-fit allocation (CA-BFD) method; the CA-BFD method firstly distributes high-key-level tasks and then distributes low-key-level tasks; adopting an optimal adaptation method to distribute tasks; 10000 task sets are generated in each experiment, and the experimental results are averaged; experimental results show that compared with the CA-BFD method, the method provided by the invention has the advantage that the energy consumption is reduced by 9.3%.

The above is only one preferred embodiment of the examples of the present invention. However, the present invention is not limited to the above embodiments, and all equivalent changes and modifications can be made according to the present invention without departing from the scope of the present invention.

Claims (5)

1. The energy-saving scheduling method for the non-precise mixed critical tasks based on the group task partition is characterized by comprising the following steps of:

step 1, scheduling a non-precise mixed critical periodic task Γ on a multiprocessor system P using a dynamic priority strategy; dividing the non-exact mixed critical task set Γ into low-level critical task sets Γ _LO And a high key hierarchy task set Γ _HI For gamma _LO According to the utilization rate in the low modeNon-ascending sort is performed on Γ _HI According to the task of high mode utilization +.>Non-ascending sort is performed;

step 2, Γ is set _HI And Γ _LO The tasks of the group task are combined into a group task according to a one-to-one correspondence, and the tasks which cannot form the group task are remained as an independent unit;

step 3, distributing high-key-level tasks for the group tasks by using an optimal adaptation method, then distributing low-key-level tasks, and distributing tasks for non-group tasks by using the optimal adaptation method;

step 4, calculating the processor P _l Energy consumption optimization speed S of (2) _l And normalizing the energy consumptionCalculating normalized energy consumption of the whole task set>Wherein P is _l Is a processor in the multiprocessor system P.

2. The energy-saving scheduling method for non-precise mixed critical tasks based on group task partitioning as claimed in claim 1, wherein said step 1 specifically comprises:

in a multiprocessor system p= { P ₁ ,P ₂ ,…,P _M Scheduling imprecise mixing using dynamic priority policies on }Critical periodic task Γ= { τ ₁ ,τ ₂ ,…,τ _n -a }; wherein M is the number of processors, and the value of M is a positive integer; each mixed critical period task τ _i From triples { T _i ,ξ _i ,C _i -composition; wherein i is more than or equal to 1 and less than or equal to n, n represents the number of tasks in the task set, i is an integer, and M is less than n; t (T) _i For mixing critical periodic tasks τ _i Is a period of (2); zeta type toy _i For mixing critical periodic tasks τ _i Is denoted as xi _i = { HI, LO }, when mixing critical cycle task τ _i Key hierarchy ζ of (a) _i When HI, it is a high-critical-level task; when mixing critical period task tau _i Key hierarchy ζ of (a) _i When=lo, it is a low critical level task; c (C) _i For mixing critical periodic tasks τ _i Worst-case execution times in different modes; c (C) _i (HI) and C _i (LO) are respectively mixed critical period tasks τ _i Worst-case execution times in high mode and low mode; high pattern means that there is a high critical hierarchy task τ _i At processor speed S, execution times all exceed C _i (LO)/S; low mode refers to any task τ _i At processor speed S, execution time does not exceed C _i (LO)/S; if mixing critical period task tau _i For high key hierarchy task, then C _i (HI)≥C _i (LO); if mixing critical period task tau _i For low critical level tasks, then C _i (HI)≤C _i (LO); dividing the non-exact mixed critical task set Γ into low-level critical task sets Γ _LO And a high key hierarchy task set Γ _HI ；And->Respectively, imprecise mixing critical tasks tau _i Utilization in low mode and high mode; to gamma _LO According to the utilization in low mode +.>Non-ascending sort is performed on Γ _HI According to the task of high mode utilization +.>Non-ascending sort is performed.

3. The energy-saving scheduling method for non-precise mixed critical tasks based on group task partitioning according to claim 2, wherein the step 2 specifically comprises:

will be gamma _HI First high-critical-level task and Γ in (1) _LO The first low-key-level task in (1) is combined into a group task, and Γ is calculated _HI The second high-critical-level task and Γ in (1) _LO The second low-key-level task in (1) is combined into a group task, and Γ is calculated _HI Third high-critical-level task and Γ in (1) _LO The third low-key-level task in the list is combined into a group task until only tasks which do not form the group task remain, and the tasks which do not form the group task remain as a separate unit.

4. The energy-saving scheduling method for non-precise mixed critical tasks based on group task partitioning as claimed in claim 3, wherein said step 3 specifically comprises:

processor P in multiprocessor system P _l Is the remaining utilization ratio U of (2) _l Is by having been allocated to the processor P _l Tasks and waits to be allocated to processor P _l Task τ _i Determined, U _l The calculation formula of (2) is as follows:

wherein U is _l Representing processor P _l Is used for the residual utilization rate of the (a);and->Respectively represent task tau _i And has been allocated to the processor P _l The sum of the utilization rates of the above tasks in the low mode and the high mode; zeta type toy _i =lo and ζ _i =hi represents task τ respectively _i The method comprises the steps of providing a low-key-level task and a high-key-level task; when processor P _l Is the remaining utilization ratio U of (2) _l Parameter x which is equal to or greater than 0 and is feasible to schedule _lo And x _up Satisfy the relation 0 < x _lo ≤x _up When less than or equal to 1, the data is distributed to the processor P _l Task set scheduling of (1) is feasible when the processor P _l Is the remaining utilization ratio U of (2) _l < 0 or the relation 0 < x _lo ≤x _up When 1 is not established, the data is distributed to the processor P _l Is not feasible for task set scheduling; for group tasks, the best adaptation method is firstly used for carrying out the high-key-level tasks xi _i =hi assigned to processor P _k Judging low key level task xi _j =lo assigned to processor P _k If the task is feasible, the task is distributed to other processors meeting the scheduling conditions by using an optimal adaptation method, and if the task is not feasible, the task is distributed to the processors meeting the scheduling conditions by using the optimal adaptation method;

processor P _l Parameter x of scheduling feasibility of (a) _up The calculation formula of (2) is as follows:

wherein,representing processor P _l The sum of the utilization of the task set phi (l) in the high mode; />Representing processor P _l The sum of the utilization of the low critical level tasks in the task set phi (l) in the low-high mode; />Representing processor P _l The sum of the utilization of the low critical level tasks in the task set phi (l) in the high mode;

processor P _l Schedule feasible parameter x of (2) _lo The calculation formula of (2) is as follows:

wherein,representing processor P _l The sum of the utilization of the high critical level tasks in the task set phi (l) in the low mode.

5. The energy-saving scheduling method for non-precise mixed critical tasks based on group task partitioning as claimed in claim 4, wherein said step 4 specifically comprises:

processor P _l Energy consumption optimization speed S of (2) _l The calculation formula is as follows:

wherein,representing processor P _l An upper bound on the utilization of task set phi (l) in low mode; />Representing processor P _l Upper bound of utilization of task set phi (l) in high mode；S _crit Representing processor P _l Key speed of (2); processor P _l The upper bound of the utilization of the task set phi (l) in low mode +.>The calculation formula of (2) is as follows:

wherein x is a deadline parameter, and the calculation formula is as follows:

processor P _l Upper bound of utilization of task set phi (l) in high modeThe calculation formula of (2) is as follows:

processor P _l Critical speed S of (2) _crit The calculation formula of (2) is as follows:

wherein P is _ind Representing frequency independent active power; m represents a dynamic power index, which is a system correlation constant, and is usually more than or equal to 2; c (C) _ef Representing the effective switched capacitance;

computing processor P _l Normalized energy consumption of (2)The formula of (2) is as follows:

calculating normalized energy consumption of the whole task setThe formula of (2) is as follows: