CN106773711B - A kind of the hybrid tasks scheduling method and model of railway locomotive operation steerable system - Google Patents
A kind of the hybrid tasks scheduling method and model of railway locomotive operation steerable system Download PDFInfo
- Publication number
- CN106773711B CN106773711B CN201710025822.9A CN201710025822A CN106773711B CN 106773711 B CN106773711 B CN 106773711B CN 201710025822 A CN201710025822 A CN 201710025822A CN 106773711 B CN106773711 B CN 106773711B
- Authority
- CN
- China
- Prior art keywords
- task
- time
- real
- priority
- scheduling
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
- G05B13/042—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance
Abstract
The present invention provides the hybrid tasks scheduling methods and model of a kind of railway locomotive operation steerable system.Dispatching method of the invention is using a frame period as basic scheduling unit, including pretreatment;Sequence permutation is scheduled based on the second level priority rule of table- driven for periodical real-time task;It applies for aperiodicity real-time task and is arranged based on illumination scan and the method for fuzzy control;Then the recycling of timeslice and the collecting work of scheduling result and feedback data are carried out, and judges the remaining execution time;If cycle time is not run out, un-real time job is executed.Hybrid tasks scheduling method of the invention and model can greatly reduce overhead, and can compare the various change situation being setup flexibly in system implementation, and can reduce influence of the systematic uncertainty to scheduling.
Description
Technical field
The present invention relates to railway locomotive operation steerable system fields more particularly to a kind of railway locomotive to run steerable system
Hybrid tasks scheduling method and model.
Background technique
Railway locomotive operation steerable system is that it is logical to be integrated with LKJ based on Train Operation for Saving Energy mechanism and moving model
Letter, locomotive traction calculates and kinematical equation, and optimization calculates, and oil consumption calculates, self adaptive control, display and some other auxiliary function
The embedded computer system of energy.It can simulate the dynamic process of train operation, input parameter perception column according to LKJ
The variation such as vehicle, route and environment calculates velocity-distance graph and gear-suitable for locomotive apart from optimal energy conservation manipulation plan
Slightly, and optimal control is provided in very short time interval.Since system operates in the environment of control vehicle in real time, so system is appointed
Be sure to need in real time, accurately obtain completion.For the rational management of task, system delay is reduced to safeguards system normal table
Operation has vital effect.
It is a typical complex optimization control system that railway locomotive, which runs steerable system, and task model is by periodicity
The mixed task set that real-time task, aperiodicity real-time task and un-real time job collectively form.Although current computer and embedding
Enter the development of formula development technique quickly, the processing capacity of hardware device is more and more stronger, but computing resource remains most valuable, excellent
The real-time task scheduling problem of networked control systems is always the hot and difficult issue of computer field research.
The dispatching method of existing mixed task set can be divided into according to the difference of driving method: table- driven, priority are driven
Dynamic, search driving etc..
The Real-Time Scheduling scheme of table- driven is the scheduling scheme based on static scheduling table algorithm, in space shuttle software
Extensive application in the High Definition Systems such as system, missile guidance system.This scheduling scheme schedules expense is small, running precision
Height has strong guarantee for the scheduling of hard real time task.But the disadvantage is that inflexible, can only complete to the period
Property task scheduling, and need the arrival time of accurate position each task in one cycle, ready time and the worst execution
Time.It exists simultaneously in the Task Scheduling Model of railway locomotive operation steerable system and periodically appoints in real time with aperiodicity
Business, aperiodicity task reach at random, and periodic task also can be with system state change and dynamic changes.So this
Scheduling scheme can not be competent at this uncertain dispatch environment.
The Real-Time Scheduling scheme of priority driven can be subdivided into static priority scheduling and dynamic priority scheduling again.It is
A kind of very common scheduling scheme, applied to classical real time operating system VxWorks, in μ C/OS, scheduling scheme have compared with
Strong flexibility.But its flexibility is built upon on the basis of frequent task switching, needs to consume a large amount of system resources.
And in railway locomotive operation steerable system, the priority of control task will be apparently higher than optimization task, periodically complete control
Process is the most fundamental demand of system.If the scheduling scheme seized using priority, will cause: first, periodically control is appointed
Business can not be timely completed due to pirority inversion etc.;Second, the frequent switching of context causes resource utilization low, in height
System is unstable under loading condition.So the scheduling scheme based on priority driven can not solve railway locomotive operation manipulation system
The scheduling problem of system.
The Real-Time Scheduling scheme of search driving common are myopic algorithm and saving algrithm.The scheduling scheme has stronger
Flexibility, but a disadvantage is that the expense of search process is larger, for railway locomotive operation steerable system, using the side of search
The success rate of algorithms for periodic task scheduling can be greatly reduced in property dispatching cycle task and aperiodicity task to formula simultaneously, thus serious shadow
Acoustic system performance.
So existing various scheduling schemes can not all adapt to the dispatching requirement in complex optimization control system at present.If
The real-time of control and the stability of optimization and the tune that can give full play to processor maximum performance can be met simultaneously by counting out one
Degree scheme is most important.
Summary of the invention
The object of the present invention is to provide the hybrid tasks scheduling methods and model of a kind of railway locomotive operation steerable system.Iron
The task-set of road locomotive system is the hybrid task collectively formed by periodic task, aperiodicity task and un-real time job
Collection.Wherein periodic task is the core of entire control process, has most stringent of time constraints, executes period diversification, when
Sequence is strong, and resource-sharing and mutual exclusion situation are more;Aperiodicity task is that maintenance system is stablized, and improves system performance and system optimization
Important component, have it is random reach, task type and form are abundant, the big feature of priority variation range;It is non real-time
Task is the guarantee of system maintenance and debugging, without the requirement handled in real time, but is calculated time-consuming maximum.The present invention is not in view of
The dispatching requirement of same type task can clearly describe the time-constrain of task itself, temporal constraint and resource between task
The characteristics of constraining, accurately embodying task execution in actual scene, realizes the rational management of hybrid task.
The present invention is achieved through the following technical solutions:
A kind of hybrid tasks scheduling method of railway locomotive operation steerable system, which is characterized in that the hybrid task packet
It includes: periodical real-time task, aperiodicity real-time task and un-real time job;The hybrid tasks scheduling method is with a frame
Period includes the following steps: as basic scheduling unit, the hybrid tasks scheduling method
(1) when the frame period starts, the hybrid tasks scheduling method is scheduled preparation, and the scheduling preparation includes at least
Task switching and resource inspection;
(2) for periodical real-time task, it is scheduled sequence permutation using the second level priority rule based on table- driven,
And according to the sequence property dispatching cycle real-time task;
(3) it is directed to aperiodicity real-time task, using the method based on illumination scan and fuzzy control, to non-week
Phase property real-time task is arranged, and dispatches aperiodicity real-time task according to the sequence;
(4) result is recycled and be scheduled to timeslice not used in current period and feedback data is adopted
Collection judges the remaining execution time;
(5) if current period time is not run out, un-real time job is executed;
(6) in frame end cycle, if aperiodicity real-time task or un-real time job are also not carried out end, under
Periodical real-time task is executed when one frame period started, and seals the aperiodicity real-time task for being not carried out completion or non real-time up for safekeeping
Business, it is ensured that the real-time of periodical real-time task.
Further, the second level priority rule based on table- driven is to transport for periodical real-time task in locomotive
A kind of non-preemption dispatching method that predictability in row steerable system proposes, the second level priority be divided into group priority and
Example priority first by group priority ranking in the step (2), then presses example priority ranking, ultimately generates dispatch list.
Further, the method for introducing fuzzy control based on illumination scan proposes a kind of new
Evaluation function:
H (n, t)=h (n, t) * nst
nst=(nd-t)-(nc-e(n,t))
Wherein H (n, t) is evaluation function, and h (n, t) is dynamic threshold coefficient of the aperiodicity real-time task n in moment t,
nstIt is remaining free time of the aperiodicity real-time task n in moment t, ndIt is the off period of task n, ncWhen being the worst execution
Between, e (n, t) is the time that task n had been executed in moment t.
Further, in the second level priority rule described based on table- driven, the design rule of described group of priority are as follows:
Sequence first ascending according to the off period of real-time task periodical in task-set, is arranged several task groups, then basis
Temporal constraint between task determines the predecessor task of each periodic task and adjusts corresponding task groups priority;Described
Example priority is the ordering rule of task groups internal task, and sequence is using the worst minimum execution time rule.
Further, in the evaluation function, the completion rate of aperiodicity real-time task is also used as Indistinct Input parameter,
It inputs parameter and threshold coefficient mapping table is as shown in the table:
Task completion rate is defined as follows: ncr(t)=e (n, t)/nc。
Further, in the step (5), for un-real time job, dispatching sequence is scheduled according to FIFO rule.
On the other hand, the present invention provides a kind of hybrid tasks scheduling model of railway locomotive operation steerable system, feature
It is, the scheduling model is divided into two parts, respectively scheduling unit and execution unit, and the scheduling unit is wanted using right
Method described in any one of 1-6 is asked to be scheduled hybrid task.
Further, the scheduling unit includes task collector, real-time scheduler and task integral unit, in locomotive
In the operational process for running steerable system, constantly there is real-time task generation, the task collector is used for generated task
It is focused on, when each clock signal arrives, submits to real-time scheduler;The real-time scheduler is in next cycle
Before arrival, periodical Real-Time Task Schedule Algorithm and aperiodicity real-time task scheduling is called to calculate respectively for different type task
Method, it is corresponding to generate periodical real-time task scheduling sequence and aperiodicity real-time task scheduling sequence;Finally by the task
Integral unit, which merges two sequences, generates a United Dispatching sequence, submits to the task executing units;It is described to execute list
Member executes task schedule according to schedule sequences, and scheduling result and intermediate state information are fed back to the real-time scheduler, by
The real-time scheduler is acquired and saves.
Dynamic threshold coefficient mentioned in the present invention refers to multiplying by fuzzy control method in assessment values in the present invention
With a parameter, this parameter is determined by fuzzy control method, different values is had in the state of difference, referred to here as it
For dynamic threshold coefficient.
Beneficial effect by adopting the above technical scheme is:
(1) algorithm of the invention for periodical real-time task design, can be greatly compared to priority scheduling algorithm
Reduce overhead, expense and processor including task switching execute the expense of dispatching algorithm;Compared to static table driven
Dispatching algorithm, the algorithm can compare the various change situation being setup flexibly in system implementation again.And it is steady in system
After fixed operation, change expense is very small required for the dispatch list that the method for the present invention generates.
(2) present invention employs a kind of evaluation functions of brand-new design, using fuzzy control method to the ginseng of evaluation function
Number is just blurred, and on the basis of guaranteeing to minimize the aperiodicity real-time task response time, mainly there is 3 benefits: the
One, fuzzy application can reduce influence of the systematic uncertainty to scheduling, and in actual operation, the state of system is can not essence
It really predicts, specifies task parameters different from scheduling model, the time parameter of task can have small under system different conditions
Difference, and fuzzy technology can weaken this difference bring and influence;Second, fuzzy theory, which can integrate, multifactor to be worked as task
Task free time and completion rate in the cumulative influence of preceding state, such as present invention design;Third, Fuzzy Threshold can reduce
The shake of rim condition dispatching algorithm, if it is the judgement of fixed value, system is easy to generate and jolt under boundary condition.
Detailed description of the invention
Fig. 1 is the hybrid tasks scheduling method flow diagram that railway locomotive of the invention runs steerable system;
Fig. 2 is the task schedule precedence diagram of dispatching method of the invention within a frame period;
Fig. 3 A and Fig. 3 B are the present invention when carrying out aperiodicity real-time task, free time and completion rate subordinating degree function
Relational graph;
Fig. 4 is the hybrid tasks scheduling model schematic that railway locomotive of the invention runs steerable system.
Specific embodiment
To keep the present invention relatively sharp, the present invention is described in detail with reference to the accompanying drawing.
Firstly, running all kinds of set of tasks in steerable system for clear description railway locomotive, the present invention defines task mould
Type is as follows:
(1) periodical real-time task: periodical real-time task is the real-time task that there is fixed cycle, circulation to execute.
Periodical real-time task can be expressed as to a five-tuple:
P=(a, r, t, d, c)
The element of tuple is respectively as follows: the arrival time of periodical real-time task, the ready time of periodical real-time task, week
On the uniprocessor of off period, periodical real-time task that the execution period of phase property real-time task, periodical real-time task execute
The worst execution time.
(2) aperiodicity real-time task: aperiodicity real-time task is arrival time random real-time task, while it
Execute the time be it is uncertain, aperiodicity real-time task can be expressed as to a four-tuple:
N=(a, r, d, c)
The element of tuple be respectively as follows: the arrival time of aperiodicity real-time task, aperiodicity real-time task it is ready when
Between, aperiodicity real-time task execute off period, aperiodicity real-time task uniprocessor on the worst execution time.
(3) un-real time job: un-real time job is other tasks in system in addition to real-time task, with a four-tuple
To indicate:
O=(a, r, s, l)
The element of tuple is respectively as follows: the arrival time of un-real time job, the ready time of un-real time job, un-real time job
The delay time of Starting Executing Time, un-real time job.
The present embodiment provides a kind of hybrid tasks scheduling methods of railway locomotive operation steerable system, and specific implementation process is such as
Shown in Fig. 1 comprising:
Step S101, when the frame period starts, dispatching method is prepared work, including task switching, resource inspection etc..
Dispatching method of the invention is using a frame period as basic scheduling unit, and wherein the frame period is the base of system operation
Describe the minimum unit of system period divisions paracycle.The duty cycle that periodical real-time task is concentrated is with times in frame period
Number relationship is described.
As shown in Fig. 2, dispatching method needs task when carrying out last end cycle to switch when a frame period starts
Work, the task that when usual last frame end cycle may be carrying out are aperiodicity real-time task or un-real time job, with
Upper two kinds of task priority in railway locomotive operation steerable system of the invention is lower than periodical real-time task.Simultaneously
Algorithm needs to carry out resource inspection, it is ensured that the task in last period has released related resource.
Step S102 is scheduled sequence using the second level priority rule based on table- driven for periodical real-time task
Column sequence, and according to the sequence property dispatching cycle real-time task.
Periodical real-time task is most important part in locomotive operation steerable system hybrid task, is usually carry in real time
The function of control, so should also distribute higher priority.Although locomotive operation steerable system is a complicated optimization control
System processed, once but the parameter of system determines that then the time parameter of periodical real-time task is also predictable.Meanwhile
Under different application scenarios and environment, the operating status of system is different, and periodical real-time task can also occur dynamically to change.
In order to make full use of predictability of the periodical real-time task in control process, while targetedly handling
Uncertainty in the implementation procedure, the present invention devise a kind of non-preemption scheduling of second level priority rule based on table- driven
Algorithm.Temporal constraint relationship between off period and task that the second-level dispatching rule investigates each task.When it is implemented, each
There are two priority, respectively group priority and example priority for periodical real-time task tool.In the tune of periodical real-time task
When degree series arrange, first by group priority ranking, then example priority ranking is pressed, ultimately generate schedule sequences.
Specific create-rule is as follows: (1) in order to make full use of the utilization rate of processor, first, in accordance with the off period in task-set
Several task groups are arranged in ascending sequence.According to the period of task, each periodical real-time task is assigned to task
In group, having the same group of priority of identical group of periodic task, i.e., significance level having the same;(2) according between task
Temporal constraint, determine the predecessor task of each periodic task.It should if the group priority of the predecessor task of the task is greater than
Task, then maintaining the group priority of two current tasks constant;Otherwise, the predecessor task of the task is redistributed to this
In the priority group of business;(3) use the worst minimum execution time as criterion the priority orders of task in organizing, i.e.,
The worst time smaller task that executes has higher example priority.After setting up the example priority of whole tasks, then root
According to the relationship of temporal constraint, task instances priority orders are adjusted, temporal constraint can be met;(4) task is determined
Execute sequence, first according to a group priority ranking, then task identical for group priority, from high to low according to example priority
It is ranked up, ultimately generates algorithms for periodic task scheduling table.
Second level priority rule has taken into account the dynamic variation characteristic of the off period of computing cost and task.By utilizing the period
The predictability of property real-time task can do some pretreatments in system initialisation phase to the parameter of dispatch list, such as preferentially
The setting of grade group, the pre-generatmg of dispatch list.In actual system design process, since periodical real-time task has centainly
Local time's characteristic, i.e., whithin a period of time, periodical real-time task collection is stable.So being generated for the first time in dispatch list
Afterwards, later each generation can modify on the basis of upper primary generation, and required change expense is very little.
Step S103, for aperiodicity real-time task, using the method based on illumination scan and fuzzy control,
Aperiodicity real-time task is arranged, and according to the sequence scheduler task.
The characteristics of aperiodicity real-time task is to reach at random, when the arrival of the unpredictable aperiodicity real-time task of system
Between.Regulation goal for aperiodicity real-time task is: meeting the premise of off period in the periodical real-time task scheduling of guarantee
Under, the completion success rate of aperiodicity real-time task is maximized, while making the response time of aperiodicity real-time task minimum.
The characteristics of integrated system regulation goal and mixed task set, aperiodicity real-time task scheduling problem can be regarded as
A kind of constrained route searching problem, i.e., arranged rational aperiodicity is appointed in real time in periodical real-time task remaining time
Business, allows aperiodicity real-time task scheduling success rate and the utilization rate of processor all to reach comparatively ideal value.The present invention devises
A kind of aperiodicity Real-Time Task Schedule Algorithm based on illumination scan, and with the thought of fuzzy control, it reduces and calculates
The unstability of method.
Illumination scan process are as follows: for the state space E being made of ordered n-tuple group, give about in n tuple points
One constraint set D and optimal objective function C (x) of amount are quickly obtained full in E by assessing each searching position
A n tuple of whole constraint conditions in foot constraint D, which is exactly the optimal solution of heuristic search problems.In the present invention,
The n tuple corresponds to the schedule sequences of aperiodicity real-time task, and the tuple for meeting requirement of the constraint in conjunction with D belongs to schedulable sequence
Column, if the tuple makes optimal objective function C (x) reach extreme value simultaneously, which is optimal adjustable degree series.Specific implementation
When, the greedy mode of algorithm application is selected the maximum task of assessment values every time from set of tasks by evaluation function and carries out sequence
Sequence.It can be proved that if evaluation function design rationally, meets monotonicity requirement, at the same task-set be it is schedulable, pass through this
The schedulable node that the greedy mode of kind obtains must be optimal schedulable node.
The key of illumination scan is the design of evaluation function, and present invention introduces the thoughts of fuzzy control, is proposed
A kind of new evaluation function design method.Specific design process is as follows:
(1) free time of aperiodicity real-time task is defined are as follows:
nst=(nd-t)-(nc-e(n,t))
Define the completion rate of aperiodicity real-time task are as follows:
ncr(t)=e (n, t)/nc
Wherein, nstIt is remaining free time of the aperiodicity real-time task n in moment t, ndIt is the off period of task n, nc
It is the worst execution time, e (n, t) is the time that task n had been executed in moment t.
If the free time of an aperiodicity real-time task is smaller, it is more urgent, more should priority scheduling;Such as
The completion rate of one aperiodicity real-time task of fruit is higher, then it more needs to stay in the processor more, goes to complete remaining portion
Point, it needs preferentially to be dispatched.
(2) present invention devises a kind of Fuzzy Threshold rule, can integrate aperiodicity real-time task free time and
Completion rate determines its significance level.The free time of aperiodicity real-time task and completion rate subordinating degree function such as Fig. 3 A and figure
Shown in 3B.
Different inputs are corresponded to three linguistic fuzzy sets respectively, free time: it is short, in, long and completion rate: it is low, in,
It is high.Complicated in order to avoid calculating during real-time control, the present invention represents the Fuzzy Threshold of task using specific output
Coefficient, during real-time calculate, by tabling look-up, to obtain the time complexity of the Fuzzy Threshold of aperiodicity real-time task be O
(1).The threshold coefficient mapping table of Indistinct Input is as shown in the table:
The first row specifies aperiodicity real-time task free time grade in above-mentioned relation table, i.e., current task is urgent
Situation, first row specify the completion rate of aperiodicity real-time task, i.e. the current task urgent degree that needs to continue to complete.Tool
Current free time and completion rate of all unscheduled aperiodic real-time tasks in moment t is first calculated when body is implemented.According to sky
Degree of membership corresponding with completion rate calculating finds out dynamic threshold coefficient then further according to the corresponding relationship in table in proportion between idle.
(3) it is as follows to define evaluation function by the present invention:
H (n, t)=h (n, t) * nst
Wherein H (n, t) is evaluation function of the aperiodicity real-time task in moment t, and h (n, t) is that aperiodicity is appointed in real time
N be engaged in the dynamic threshold of moment t, the calculation method of the value calculated by the statement in evaluation function design process (2) step.
The schedule sequences sequence that aperiodicity real-time task can be realized by above step, can be appointed according to the sequence
The scheduling process of business.
Step S104, post-processing carries out the recycling of timeslice and the collecting work of scheduling result and feedback data, and sentences
The disconnected remaining execution time.
The task schedule sequence of dispatching algorithm of the invention within a frame period is as shown in Figure 2.Step S102 and S103
Achievable hybrid task concentrates the schedule sequences sequence of periodical real-time task and aperiodicity real-time task, and completes to dispatch
Journey.The recycling of timeslice and the receipts of scheduling result and feedback data in the frame period are carried out followed by algorithm last handling process
Collect work, the above work is the time resource made full use of in the frame period, and standard has been made in the scheduling of further progress un-real time job
It is standby.
Step S105 executes un-real time job if cycle time is not run out, dispatching sequence according to FIFO rule into
Row scheduling.
Investigate that step S104 is post-processed as a result, appointing if still there is remaining timeslice in current frame period from mixing
Business concentrates selection non real-time tasks to be executed, and the rule of selection is FIFO rule.To make full use of in the frame period when
Between piece resource.
If current hybrid task concentrates real-time task nothing but, enter idle state.
Step S106, in frame end cycle, if aperiodicity real-time task or un-real time job are also not carried out end,
It can then be seized by the periodical real-time task in next frame period, it is ensured that the real-time of periodical real-time task.
It is executed by all scheduling that mixed task set in system can be realized in step S101 to step S105.It is tied in the period
Shu Shi, it is understood that there may be aperiodicity real-time task or it is non real-time still execute the case where, the two kinds of task is preferential at this time
Grade is lower than periodical real-time task, then can be robbed in next frame period by the frame period real-time task in next frame period
It accounts for, so that it is guaranteed that the real-time of periodical real-time task.
The above S101 to S106 is the execution step of task scheduling algorithm in a frame period, within next frame period still
The scheduling of mixed task set is executed according to algorithm sequence.
The hybrid tasks scheduling model of railway locomotive operation operating system proposed by the present invention is as shown in figure 4, include task
Scheduling unit and execution unit.Scheduling unit includes task collector, real-time scheduler and task integral unit.It is run in system
In the process, constantly there is real-time task generation, source includes human-computer interaction, outer signals, system feedback etc..Task initialization
Afterwards, it is focused on by task collector, when each clock signal arrives, submits to real-time scheduler;Real-time scheduler is under
Before a cycle arrives, integrate these tasks, is called respectively for different type task periodical Real-Time Task Schedule Algorithm with
Aperiodicity Real-Time Task Schedule Algorithm, it is corresponding to generate periodical real-time task scheduling sequence and aperiodicity real-time task scheduling
Sequence;Module is integrated finally by task, two sequences are merged into one United Dispatching sequence of generation, submit to task execution list
Member.Execution unit is scheduled task according to the instruction of schedule sequences, and scheduling result and some intermediate state are fed back to
Real-time scheduler is acquired by scheduler and is saved.
The present invention devises a kind of based on table for periodical real-time task it can be seen from the technical solution of foregoing invention
The non-preemption dispatching algorithm of the second level priority rule of driving has lower overhead, has combined priority scheduling
Flexibility;A kind of aperiodicity real-time task tune based on illumination scan is devised for aperiodicity real-time task
Algorithm is spent, feasibility analysis is used to task-set, limits the depth of recursion of heuristic search algorithm, heuristic evaluation function is joined
Digital-to-analogue gelatinization, reduces the shake generated when scheduling, which can preferably handle aperiodicity real-time task scheduling;For non-
Real-time task takes full advantage of the time resource and computing resource of system, farthest guarantees with classical FIFO rule
Un-real time job is executed by time scheduling.The method and model that the present invention designs can effectively solve complicated railway locomotive operation
The hybrid tasks scheduling demand of steerable system.
Although the principle of the present invention is described in detail above in conjunction with the preferred embodiment of the present invention, this field skill
Art personnel are it should be understood that above-described embodiment is only the explanation to exemplary implementation of the invention, not to present invention packet
Restriction containing range.Details in embodiment is simultaneously not meant to limit the scope of the invention, without departing substantially from spirit of the invention and
In the case where range, any equivalent transformation based on technical solution of the present invention, simple replacement etc. obviously change, and all fall within
Within the scope of the present invention.
Claims (7)
1. a kind of hybrid tasks scheduling method of railway locomotive operation steerable system, which is characterized in that the hybrid task includes:
Periodical real-time task, aperiodicity real-time task and un-real time job;The hybrid tasks scheduling method is with a frame week
Phase includes the following steps: as basic scheduling unit, the hybrid tasks scheduling method
(1) when the frame period starts, the hybrid tasks scheduling method is scheduled preparation, and the scheduling prepares to include at least task
Switching and resource inspection;
(2) for periodical real-time task, it is scheduled sequence permutation using the second level priority rule based on table- driven, and press
According to the sequence property dispatching cycle real-time task, between off period and task which investigates each task
Temporal constraint relationship, when it is implemented, there are two priority, respectively group priority and example for each periodicity real-time task tool
Priority first by group priority ranking, then presses example priority ranking in the schedule sequences arrangement of periodical real-time task,
Ultimately generate schedule sequences,
Wherein, second level priority scheduling rules are as follows:
(2.1) in order to make full use of the utilization rate of processor, first, in accordance with the sequence that the off period in task-set is ascending, setting
Each periodical real-time task is assigned in task groups, identical group of periodicity by several task groups according to the period of task
Having the same group of priority of task, i.e., significance level having the same;(2.2) according to the temporal constraint between task, determine every
The predecessor task of a periodic task remains current if the group priority of the predecessor task of the task is greater than the task
Two tasks group priority it is constant;Otherwise, the predecessor task of the task is redistributed in the priority group of the task;
(2.3) use the worst minimum execution time as criterion, i.e., the worst execution time the priority orders of task in organizing
Smaller task has higher example priority, after setting up the example priority of whole tasks, further according to temporal constraint
Relationship adjusts task instances priority orders, temporal constraint can be met;(2.4) sequence that executes of task is determined, first
According to a group priority ranking, then task identical for group priority, is ranked up, most from high to low according to example priority
Throughout one's life at algorithms for periodic task scheduling table;
(3) it is directed to aperiodicity real-time task, using the method based on illumination scan and fuzzy control, to aperiodicity
Real-time task is arranged, and dispatches aperiodicity real-time task according to the sequence, wherein described based on heuristic search plan
The method for slightly introducing fuzzy control proposes a kind of new evaluation function:
H (n, t)=h (n, t) * nst
nst=(nd-t)-(nc- e (n, t))
Wherein H (n, t) is evaluation function, and h (n, t) is dynamic threshold coefficient of the aperiodicity real-time task n in moment t, nstIt is
Remaining free time of the aperiodicity real-time task n in moment t, ndIt is the off period of task n, ncIt is the worst execution time, e (n,
T) it is time that task n had been executed in moment t;
(4) acquisition for timeslice not used in current period being recycled and being scheduled result and feedback data, sentences
The disconnected remaining execution time;
(5) if current period time is not run out, un-real time job is executed;
(6) in frame end cycle, if aperiodicity real-time task or un-real time job are also not carried out end, in next frame
Periodical real-time task is executed when period starts, and seals the aperiodicity real-time task or un-real time job for being not carried out completion up for safekeeping, really
Protect the real-time of periodical real-time task.
2. a kind of hybrid tasks scheduling method of railway locomotive operation steerable system according to claim 1, feature exist
In the second level priority rule based on table- driven is for periodical real-time task in locomotive operation steerable system
A kind of non-preemption dispatching method that predictability proposes, the second level priority are divided into group priority and example priority,
First by group priority ranking in the step (2), then example priority ranking is pressed, ultimately generates dispatch list.
3. a kind of hybrid tasks scheduling method of railway locomotive operation steerable system according to claim 2, feature exist
In, in the second level priority rule based on table- driven, the design rule of described group of priority are as follows: first according in task-set
The off period of periodical real-time task ascending sequence, is arranged several task groups, then about according to the timing between task
Beam determines the predecessor task of each periodic task and adjusts corresponding task groups priority;The example priority is to appoint
The ordering rule of business group internal task, sequence is using the worst minimum execution time rule.
4. a kind of hybrid tasks scheduling method of railway locomotive operation steerable system according to claim 1, feature is also
It is, in the evaluation function, the completion rate of aperiodicity real-time task is also used as Indistinct Input parameter, inputs parameter and threshold
Value coefficient mapping table is as shown in the table:
Task completion rate is defined as follows: ncr(t)=e (n, t)/nc。
5. a kind of hybrid tasks scheduling method of railway locomotive operation steerable system according to claim 1, feature is also
It is, in the step (5), for un-real time job, dispatching sequence is scheduled according to FIFO rule.
6. a kind of hybrid tasks scheduling model of railway locomotive operation steerable system, which is characterized in that the scheduling model point
For two parts, respectively scheduling unit and execution unit, the scheduling unit is using described in any one of claim 1-5
Method is scheduled hybrid task.
7. a kind of hybrid tasks scheduling model of railway locomotive operation steerable system according to claim 6, feature exist
In the scheduling unit includes task collector, real-time scheduler and task integral unit, in locomotive operation steerable system
In operational process, constantly there is real-time task generation, the task collector is used to focus on generated task,
When each clock signal arrives, real-time scheduler is submitted to;The real-time scheduler is before next cycle arrival, for difference
Type tasks call periodical Real-Time Task Schedule Algorithm and aperiodicity Real-Time Task Schedule Algorithm respectively, corresponding to generate the period
Property real-time task scheduling sequence and aperiodicity real-time task scheduling sequence;Finally by the task integral unit by two sequences
Column, which merge, generates a United Dispatching sequence, submits to the execution unit;The execution unit is executed according to schedule sequences appoints
Business scheduling, and scheduling result and intermediate state information are fed back into the real-time scheduler, it is adopted by the real-time scheduler
Collection and preservation.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710025822.9A CN106773711B (en) | 2017-01-13 | 2017-01-13 | A kind of the hybrid tasks scheduling method and model of railway locomotive operation steerable system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710025822.9A CN106773711B (en) | 2017-01-13 | 2017-01-13 | A kind of the hybrid tasks scheduling method and model of railway locomotive operation steerable system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106773711A CN106773711A (en) | 2017-05-31 |
CN106773711B true CN106773711B (en) | 2019-09-17 |
Family
ID=58946467
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710025822.9A Active CN106773711B (en) | 2017-01-13 | 2017-01-13 | A kind of the hybrid tasks scheduling method and model of railway locomotive operation steerable system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106773711B (en) |
Families Citing this family (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108733473A (en) * | 2018-05-11 | 2018-11-02 | 北京航天发射技术研究所 | A kind of positioning based on VxWorks aims at the control method of integration apparatus task |
US10691490B2 (en) * | 2018-07-06 | 2020-06-23 | Apple Inc. | System for scheduling threads for execution |
WO2020252618A1 (en) * | 2019-06-17 | 2020-12-24 | Beijing Voyager Technology Co., Ltd. | Systems and methods for controlling autonomous vehicle in real time |
CN112214291A (en) * | 2019-07-12 | 2021-01-12 | 杭州海康汽车技术有限公司 | Task scheduling method and device |
CN110620735B (en) * | 2019-09-09 | 2023-02-28 | 北京航天自动控制研究所 | Dynamic load balancing telemetering method based on tabular configuration |
CN111314801B (en) * | 2020-02-13 | 2022-01-28 | 中国铁道科学研究院集团有限公司铁道建筑研究所 | Data acquisition system and method supporting dynamic scheduling |
CN113495781B (en) * | 2021-06-30 | 2023-03-03 | 东风商用车有限公司 | Task scheduling method, device, equipment and readable storage medium |
CN113419834B (en) * | 2021-07-02 | 2023-02-17 | 中国船舶重工集团公司第七二四研究所 | Multitask time slot self-adaptive heterogeneous radar resource scheduling method |
CN117132871A (en) * | 2022-05-18 | 2023-11-28 | 阿里云计算有限公司 | Model scheduling method, device and storage medium |
CN114661449B (en) * | 2022-05-19 | 2022-08-19 | 四川傲势科技有限公司 | Task scheduling method, embedded system and computer readable storage medium |
CN115793623B (en) * | 2023-02-08 | 2023-05-16 | 北京理工大学 | Hardware-in-loop testing device and method for vehicle hybrid key system |
Family Cites Families (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20060212188A1 (en) * | 2003-02-27 | 2006-09-21 | Joel Kickbusch | Method and apparatus for automatic selection of alternative routing through congested areas using congestion prediction metrics |
GB0513045D0 (en) * | 2005-06-27 | 2005-08-03 | Vidus Ltd | Resource scheduling method and system |
CN101859347A (en) * | 2010-05-26 | 2010-10-13 | 北京交通大学 | Method using intelligent hybrid model to describe operational process of high-speed train group |
US8924976B2 (en) * | 2011-08-26 | 2014-12-30 | Knu-Industry Cooperation Foundation | Task scheduling method and apparatus |
CN103377078B (en) * | 2012-04-11 | 2017-04-12 | 广州地铁集团有限公司 | Real-time task scheduling method and system for vehicular ATP |
CN105260497B (en) * | 2015-08-07 | 2019-04-30 | 黑龙江大学 | Real-time task schedulability based on linear linked list tests semi-direct analogy method |
-
2017
- 2017-01-13 CN CN201710025822.9A patent/CN106773711B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN106773711A (en) | 2017-05-31 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106773711B (en) | A kind of the hybrid tasks scheduling method and model of railway locomotive operation steerable system | |
Tang et al. | An improved differential evolution algorithm for practical dynamic scheduling in steelmaking-continuous casting production | |
CN102135906B (en) | Power consumption control method and system orienting to embedded real-time operating system | |
CN101237469B (en) | Method for optimizing multi-QoS grid workflow based on ant group algorithm | |
CN101620550A (en) | Embedded real-time scheduling method based on fuzzy multiple features of task | |
Cheng et al. | Deadline-aware MapReduce job scheduling with dynamic resource availability | |
Kalinowski et al. | Predictive-reactive strategy for real time scheduling of manufacturing systems | |
Valente et al. | Development of multi-level adaptive control and scheduling solutions for shop-floor automation in reconfigurable manufacturing systems | |
CN104239135A (en) | Hybrid task scheduling method of directed acyclic graph (DGA) based reconfigurable system | |
Chwa et al. | Physical-state-aware dynamic slack management for mixed-criticality systems | |
CN112101773A (en) | Task scheduling method and system for multi-agent system in process industry | |
Binns | Incremental rate monotonic scheduling for improved control system performance | |
CN101807065A (en) | Real-time task scheduling method of soft numerical control system | |
CN109308216B (en) | Single-core system real-time task scheduling method aiming at inaccurate calculation | |
CN101909055A (en) | Multithread Web service negotiation method based on QoS | |
Zeng et al. | A deep reinforcement learning approach to flexible job shop scheduling | |
Casagrande et al. | DeepScheduling: grid computing job scheduler based on deep reinforcement learning | |
KR102566824B1 (en) | A method for training long short term memory network and a method for minimizing energy costs using trained long short term memory network | |
Salmani et al. | A fuzzy-based multi-criteria scheduler for uniform multiprocessor real-time systems | |
Salmi et al. | Assembly Modelling and Time estimating during the early phase of Assembly Systems Design | |
He et al. | Hybrid earliest deadline first/preemption threshold scheduling for real-time systems | |
Rattanatamrong et al. | Mode transition for online scheduling of adaptive real-time systems on multiprocessors | |
Liu et al. | Energy optimization on dynamic multiple functions automotive cyber-physical systems | |
Mochocki et al. | Dynamic voltage scaling for the schedulability of jitter-constrained real-time embedded systems | |
Ulbrich et al. | Qronos: Towards quality-aware responsive real-time control systems |
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 |