Method and system for verifying effectiveness of resource scheduling scheme based on execution information

Abstract

The application discloses a resource scheduling scheme effectiveness verification method and system based on execution information, and belongs to the resource scheduling field. The method comprises the following steps: constructing a priori known time network, using a priori known accidental link and an observation link to represent an action with uncertain duration and an observation process that can obtain execution information in an execution process, obtaining time constraints that need to be met at all controlled time points by applying propagation constraints, realizing the use of execution information to deal with time uncertainty, verifying whether the resource scheduling scheme meets the dynamic controllable condition when facing time uncertainty by using a dynamic controllability detection algorithm, and then judging whether the resource scheduling scheme can be successfully executed after the execution information obtained in the execution process is used to reduce uncertainty. The method has considered the execution information that can be obtained in the execution process in the verification process, and improves the accuracy and efficiency of the verification.

Description

Technical Field

[0001] This invention belongs to the field of resource scheduling, and more specifically, relates to a method and system for verifying the effectiveness of resource scheduling schemes based on execution information. Background Technology

[0002] In target task scheduling, the start and end of each action typically need to meet a series of temporal constraints. For example, after receiving a task instruction, the personnel executing the target task must react quickly, arriving at the scene within 5 minutes in urban areas and within 10 minutes in suburban areas. These personnel are also called action personnel. Decision-makers need to verify the resource scheduling plan to determine whether the current plan meets all temporal constraints. In actual operations, the duration of some actions becomes uncertain due to traffic, case situation evolution, etc. In this case, it is necessary to verify whether the current resource scheduling plan can cope with all possible temporal uncertainties and be executed smoothly. To verify whether a given plan can cope with the uncertainty of action duration based on information that may be obtained during execution, the following two methods are currently used to verify the resource scheduling plan.

[0003] The first approach involves manual decision verification, while the second utilizes a temporal network model to represent the temporal constraints of the resource scheduling scheme and applies a dynamic controllability detection algorithm to verify the scheme's effectiveness. In the second method, a temporal network model is used to represent the temporal constraints in the resource scheduling scheme, and a dynamic controllability detection algorithm is used to verify whether the scheme can cope with uncertainty. In the temporal network, accidental links represent action processes with uncertain durations. The possible duration values ​​are within a given range, but due to unknown factors such as traffic conditions and the evolution of the case situation, the duration of the action cannot be controlled by the decision-maker but is chosen by the environment's non-determinism. Under temporal constraints, the dynamic controllability detection algorithm ensures that controlled time points in the resource scheduling scheme can be dynamically assigned subsequent controlled time points after observing the occurrence time of accidental time points that occurred in the past.

[0004] When using the two methods described above to verify the effectiveness of resource scheduling schemes, for actions with uncertain durations, the information learned during the execution of such actions that could reduce the uncertainty of their duration cannot be applied. However, applying this information could potentially make a resource scheduling scheme that was originally unable to cope with uncertainty manageable. Therefore, methods that ignore this information cannot accurately identify all effective resource scheduling schemes, impacting execution performance and efficiency. Summary of the Invention

[0005] To address the shortcomings and improvement needs of existing technologies, this invention provides a method and system for verifying the effectiveness of resource scheduling schemes based on execution information. The purpose is to reduce the reliance on manual processes and informal knowledge in existing resource scheduling methods and improve the accuracy of verifying the effectiveness of resource scheduling schemes.

[0006] To achieve the above objectives, according to one aspect of the present invention, a method for validating the effectiveness of a resource scheduling scheme based on execution information is provided, comprising: constructing a priori known temporal network, wherein the priori known temporal network includes controlled time points, general accidental time points, priori known accidental time points, controlled links, general accidental links, priori known accidental links, and observation links; wherein, a controlled link points to one of a controlled time point, a general accidental time point, or a priori known accidental time point, a general accidental link points to a general accidental time point or a priori known accidental time point, a priori known accidental link points to a priori known accidental time point, and an observation link includes an observation node, from which the observation node points to the priori known accidental link;

[0007] The start time of the target task execution action in the resource scheduling scheme and the end time of the action with a definite duration are taken as the controlled time point; the action with an uncertain duration that can obtain execution information is recorded as a priori known accidental action, and the action that cannot obtain execution information is recorded as accidental action. The end time of the priori known accidental action is taken as the priori known accidental time point, the end time of the accidental action is taken as the general accidental time point, and the time point of obtaining execution information is taken as the observation node.

[0008] The start and end points of actions with a definite duration are connected by controlled links; the start and end points of accidental actions are connected by general accidental links; the start and end points of a priori accidental actions are connected by a priori accidental links; and any two actions are connected by controlled links.

[0009] Based on the category of the obtained execution information, constraints that the execution information must satisfy are added to the prior known temporal network to update the duration of the prior known accidental action, and the dynamic controllability of the updated temporal network is judged. If the temporal network is controllable, the resource scheduling scheme is effective; otherwise, the resource scheduling scheme is invalid. The execution information is information obtained during the execution of an action with an uncertain duration that can reduce the uncertainty of the duration of the action.

[0010] Furthermore, the categories of execution information include any one of: fixed early execution information, variable early execution information, controlled delayed execution information, and occasional delayed execution information;

[0011] The fixed advance execution information refers to the fact that the personnel executing the target task know the end time of the prior known accidental action at a fixed time before the end of the accidental action;

[0012] The variable advance execution information refers to the duration of the prior known accidental action that the target task executor knows within a certain time range before the end of the prior known accidental action;

[0013] The controlled delayed execution information refers to the duration of the prior known accidental action that the target task executor learns within a certain time range after the start of the prior known accidental action, and the time point at which the execution information is learned is a controlled time point;

[0014] The term "accidental delayed execution information" refers to the information obtained by the target task executor within a certain time range after the prior known accidental action has begun, indicating that the duration of the prior known accidental action is known, and that the time point at which the execution information is obtained is an accidental time point.

[0015] Furthermore, in the aforementioned prior-known temporal network,

[0016] When variable advance execution information is obtained, the observation node is a general accidental time point, and there is a general accidental link between the observation node and the prior known accidental time point, and the length of the general accidental link is a fixed value o. - ;

[0017] When variable advance execution information is obtained, the observation node is a general accidental time point, and there is a general accidental link between the observation node and the prior known accidental time point, and the length of the general accidental link varies within the range of [o]. - o + ];

[0018] When controlled delay execution information is obtained, the observation node is a controlled time point, and the observation node and the time point at which the prior known accidental action begins are connected by a controlled link, with the length of the controlled link varying within the range of [o]. - o + ];

[0019] When accidental delayed execution information is obtained, the observation node is a general accidental time point, and the observation node and the time point at which the prior known accidental action begins are general accidental links, with the length of the general accidental link varying within the range of [o]. - o + ];

[0020] Among them, o - This represents the minimum value of the difference between the observed node and the time point connected to it, o. + This represents the maximum value of the difference between the observed node and the time point connected to it.

[0021] Furthermore, when variable advance execution information is obtained, the constraints added to the prior known temporal network are as follows:

[0022] Remove the prior known accidental link between the prior known start and end points of accidental actions;

[0023] Add a general accidental link from the starting point of the accidental action, which is known a priori to the observation node, to the observation node, and the length of the general accidental link varies within the range of [ao]. - ,bo - ], where a and b represent the minimum and maximum values ​​of the range of values ​​for the duration of a known accidental action, respectively.

[0024] Furthermore, when variable advance execution information is obtained, the constraints added to the prior known temporal network are as follows:

[0025] Remove the prior known accidental link between the prior known start and end points of accidental actions;

[0026] Add a general accidental link from the starting point of the accidental action, which is known a priori to the observation node, to the observation node, and the length of the general accidental link varies within the range of [ao]. - ,bo + ], where a and b represent the minimum and maximum values ​​of the range of values ​​for the duration of a known accidental action, respectively;

[0027] Change the general accidental link from the observation node to the a priori known accidental time point to a priori known accidental link, and add the observation node corresponding to the a priori known accidental link.

[0028] Furthermore, when controlled delay execution information is obtained, the constraints added to the prior known temporal network are as follows:

[0029] There exists at least one controlled time point, which occurs before a priori known accidental time point, and the controlled time point occurs at the latest after the observation node. An observation-controlled link is added between the observation node and the controlled time point; wherein, the observation-controlled link is a constraint that the controlled time point needs to satisfy.

[0030] Delete the link between the previously known accidental time point and the controlled time point;

[0031] The prior known accidental link between the known start time of the accidental action and the known accidental time point is modified to a general accidental link;

[0032] The controlled link length variation range between the observation node and the prior known start time of the accidental action is changed to a fixed value. - .

[0033] Furthermore, when information about accidental delayed execution is obtained, the constraints added to the prior known temporal network are as follows:

[0034] There exists at least one controlled time point, which occurs before a priori known accidental time point, and the controlled time point occurs at the latest after the observation node. An observation-controlled link is added between the observation node and the controlled time point; wherein, the observation-controlled link is a constraint that the controlled time point needs to satisfy.

[0035] Delete the link between the previously known accidental time point and the controlled time point;

[0036] The prior known accidental link between the known start time of the accidental action and the known accidental time point is modified to a general accidental link;

[0037] The controlled link length variation range between the observation node and the prior known start time of the accidental action is changed to a fixed value. + .

[0038] Furthermore, the dynamic controllability of the updated temporal network is determined, including:

[0039] If the duration of the prior known accidental action is a and b respectively, it is determined whether the prior known temporal network is dynamically controllable. If so, the resource scheduling scheme is effective; otherwise, the resource scheduling scheme is invalid. Here, a and b represent the minimum and maximum values ​​of the range of the duration of the prior known accidental action, respectively.

[0040] Wherein, the execution information obtained by the prior known accidental action is any one of fixed early execution information, variable early execution information, controlled delayed execution information, and accidental delayed execution information.

[0041] According to another aspect of the present invention, a resource scheduling scheme validity verification system based on execution information is provided, comprising:

[0042] A priori-knowable temporal network construction module is used to construct a priori-knowable temporal network, which includes controlled time points, general accidental time points, priori-knowable accidental time points, controlled links, general accidental links, priori-knowable accidental links, and observation links. Among them, controlled links point to controlled time points, general accidental time points, or priori-knowable accidental time points; general accidental links point to general accidental time points or priori-knowable accidental time points; priori-knowable accidental links point to priori-knowable accidental time points; and observation links contain observation nodes, which point to priori-knowable accidental links.

[0043] The temporal constraint module of the resource scheduling scheme is used to define the start time of the target task execution action and the end time of the action with a definite duration as controlled time points in the resource scheduling scheme. Actions with uncertain durations that can obtain execution information are categorized as a priori known accidental actions, while those that cannot obtain execution information are categorized as accidental actions. The end time of a priori known accidental action is defined as a priori known accidental time point, the end time of an accidental action is defined as a general accidental time point, and the time point of obtaining execution information is defined as an observation node. The start and end points of actions with definite durations are connected by controlled links, the start and end points of accidental actions are connected by general accidental links, the start and end points of priori known accidental actions are connected by priori known accidental links, and any two actions are connected by controlled links.

[0044] The network constraint propagation module is used to add constraints that the obtained execution information must satisfy to the prior known temporal network, based on the category of the execution information, in order to update the duration of the prior known accidental actions. The execution information refers to information learned during the execution of actions with uncertain durations that can reduce the uncertainty of the action's duration.

[0045] The dynamic controllability detection module is used to determine the dynamic controllability of the updated temporal network. If the updated temporal network is dynamically controllable, the resource scheduling scheme is effective; otherwise, the resource scheduling scheme is invalid.

[0046] Furthermore, the network constraint propagation module includes one or more of the following: a fixed early execution information constraint propagation unit, a variable early execution information propagation unit, a controlled delayed execution information propagation unit, and an occasional delayed execution information propagation unit;

[0047] The variable advance execution information constraint propagation unit is used to delete the a priori known accidental link between the a priori known accidental action's start and end points; and to add a general accidental link from the a priori known accidental action's start point to the observation node, wherein the length of the general accidental link varies within the range of [ao]. - ,bo - ];

[0048] The variable advance execution information propagation unit is used to delete the a priori known accidental link between the a priori known accidental action's start and end points; and to add a general accidental link from the a priori known accidental action's start point to the observation node, wherein the length of the general accidental link varies within the range of [ao]. - ,bo + ]; Change the general accidental link from the observation node to the a priori known accidental time point to the a priori known accidental link, and add the observation node corresponding to the a priori known accidental link;

[0049] The controlled delayed execution information propagation unit is used to add an observation-controlled link between the observation node and a controlled time point that occurs before a priori known accidental time point; wherein, the controlled time point also satisfies the condition that it occurs no later than the observation node, and the observation-controlled link is a constraint that the controlled time point must satisfy; delete the link between the priori known accidental time point and the controlled time point; modify the priori known accidental link between the priori known accidental action start time point and the priori known accidental time point to a general accidental link; and set the length of the controlled link between the observation node and the priori known accidental action start time point to a fixed value o. - ;

[0050] The accidental delayed execution information propagation unit is used to add an observation-controlled link between the observation node and a controlled time point that occurs before a priori known accidental time point; wherein, the controlled time point also satisfies the condition that it occurs no later than the observation node, and the observation-controlled link is a constraint that the controlled time point must satisfy; delete the link between the priori known accidental time point and the controlled time point; modify the priori known accidental link between the priori known accidental action start time point and the priori known accidental time point to a general accidental link; and set the length of the controlled link between the observation node and the priori known accidental action start time point to a fixed value o. + ;

[0051] Among them, o - This represents the minimum value of the difference between the observed node and the time point connected to it, o. + denoted by , represents the maximum value of the difference between the observation node and the time point connected to the observation node, and 'a' and 'b' represent the minimum and maximum values ​​of the range of values ​​for the duration of a priori known accidental action, respectively.

[0052] In summary, the above-described technical solutions conceived in this invention can achieve the following beneficial effects:

[0053] (1) The method for verifying the effectiveness of a resource scheduling scheme based on execution information of the present invention generally includes a temporal network construction process for the resource scheduling scheme and a dynamic controllability detection process for the resource scheduling scheme. In the temporal network construction process of the resource scheduling scheme, a novel a priori known temporal network is constructed. This a priori known temporal network uses a priori known accidental links and observation links to represent execution actions with uncertain durations and observation processes in which execution information can be obtained during execution. Through propagation constraints, the temporal constraints that need to be met by using execution information to cope with temporal uncertainty and all controlled time points are derived. The dynamic controllability detection algorithm verifies whether the resource scheduling scheme meets the conditions of dynamic controllability when facing temporal uncertainty, and then determines whether the resource scheduling scheme can be successfully executed after reducing uncertainty by using the execution information obtained during execution. Since the method of the present invention has considered the execution information that may be obtained during execution in the verification process, when it is applied to a real resource scheduling scheme to verify its effectiveness, almost no abnormal situations or errors will occur, thus improving the accuracy and efficiency of verification.

[0054] (2) Furthermore, the present invention provides an execution information propagation constraint that, for different types of execution information, can make the prior known temporal network no longer contain prior known observation links and observation links, and ensure that the value range of all time points in the network does not change. Through the dynamic controllability algorithm of the temporal network, the observation-controlled links in the network are converted into controlled links (the parameter r in the observation-controlled links is converted into the definite values ​​a and b). That is, after the propagation constraint, the final temporal network does not contain prior known observation links, observation links and observation-controlled links. It is possible to quickly determine whether the final network is dynamically controllable using existing dynamic controllability methods.

[0055] In summary, this invention provides a fully automated method for verifying resource scheduling schemes, enabling automatic determination of whether a given resource scheduling scheme can meet the temporal constraints of the current task requirements without relying on experts or experienced personnel. This scheme can be used to consider information obtainable during execution during the resource scheduling scheme generation process, verifying whether the resource scheduling scheme can successfully cope with uncertainty when the duration of the execution action is uncertain and the latest information can be dynamically obtained. Attached Figure Description

[0056] Figure 1 The structure of the four types of execution information provided in the embodiments of the present invention in a priori known temporal network.

[0057] Figure 2(A) shows the PKSTNU propagation constraints for fixed advance execution information provided in an embodiment of the present invention.

[0058] Figure 2(B) shows the PKSTNU propagation constraints for variable advance execution information provided in an embodiment of the present invention.

[0059] Figure 2(C) illustrates the PKSTNU propagation constraints for controlled delayed execution information provided in an embodiment of the present invention.

[0060] Figure 2(D) shows the PKSTNU propagation constraints for accidental delayed execution information provided in an embodiment of the present invention. Detailed Implementation

[0061] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0062] A method for validating the effectiveness of a resource scheduling scheme based on execution information, comprising:

[0063] A priori-knowable temporal network, PKSTNU, is constructed. PKSTNU comprises: controlled time points, general accidental time points, priori-knowable accidental time points, controlled links, general accidental links, priori-knowable accidental links, and observation links. Controlled links point to any one of the three time points (controlled time points, general accidental time points, and priori-knowable accidental time points). General accidental links point to either general accidental time points or priori-knowable accidental time points. Priori-knowable accidental links point to priori-knowable accidental time points. Observation links contain observation nodes, which point to priori-knowable accidental links. Observation nodes can be controlled time points, general accidental time points, or priori-knowable accidental time points. Each of the seven elements (three time points and four links) in the constructed priori-knowable temporal network PKSTNU can be represented as a set.

[0064] In the prior-knowable temporal network PKSTNU constructed in this invention, controlled time points represent controlled events; general accidental time points represent accidental events, i.e., the end time of certain processes with uncertain duration; prior-knowable accidental time points represent prior-knowable accidental events, i.e., accidental events whose specific occurrence time can be known in advance; and the four types of links respectively represent the constraints that the values ​​of the two points connected to the link must satisfy.

[0065] The resource (target task) scheduling scheme is expressed in the prior known temporal network, specifically including: the start time and end time of actions with a definite duration are regarded as controlled time points, and the end time of actions with an uncertain duration is an accidental time point. When an action is being executed, the end time of an action with an uncertain duration that can obtain execution information (denoted as a prior known accidental action) is a prior known accidental time point, and the end time of an action with an uncertain duration that cannot obtain execution information (denoted as an accidental action) is a general accidental time point. The time point at which execution information is obtained is regarded as an observation node. Here, execution information refers to information that can reduce the uncertainty of the duration of an action with an uncertain duration that is known before the execution of an action with an uncertain duration after receiving the task.

[0066] For actions with a definite duration, use controlled links to connect the start and end points of the action; for accidental actions, use general accidental links to connect the start and end points of the action; for a priori accidental actions, use a priori accidental links to connect the start and end points of the action; use controlled links to connect any two actions.

[0067] Based on the category of the obtained execution information, and based on the execution information propagation constraints, the constraints that the execution information needs to satisfy are added to the prior known temporal network PKSTNU, the duration of the prior known accidental actions is updated, and the updated temporal network containing the execution information is obtained. The dynamic controllability of the temporal network containing the execution information is then determined. If the updated temporal network is controllable, the resource scheduling scheme is effective; otherwise, the resource scheduling scheme is invalid.

[0068] Specifically, during the execution of an action, the execution information obtained by the contingent action is known a priori, including any one of the following: fixed-ahead, variable-ahead, controlled-delay, and contingent-delay.

[0069] like Figure 1 As shown in the figure, X and E represent the start and end times of a priori known accidental action, respectively; O represents the time point when execution information is obtained, i.e., the observation node; Y represents the controlled time point that needs to occur before the end of the priori known accidental action; r represents the length of the observed priori known accidental link; the interval [a, b] represents the minimum and maximum values ​​of the duration range of the priori known accidental action, i.e., a ≤ r ≤ b; and the interval [u, v] represents the controlled time point that needs to occur within a certain time range before the end of the priori known accidental action, i.e., u ≤ EY ≤ v. This indicates a coincidental link that can be known a priori from the observed node; ○ represents a controlled time point. Indicates a general, accidental point in time. This indicates a priori known accidental time point, o - o + These represent the time ranges within which the observed node can occur, where o - This represents the minimum value of the difference between the observed node and the time point connected to it, o. + This represents the maximum value of the difference between the observed node and the time point connected to it.

[0070] Fixed-ahead information refers to the information that an operator learns the end time of a priori foreseeable foreseeable action at a fixed time before the action's conclusion. For example, in a demolition mission, the demolition team can learn the specific demolition time after activating a countdown timer. In the priori-knowable temporal network PKSTNU, the time point (i.e., the observation node) at which the fixed-ahead information corresponds to the priori foreseeable foreseeable action obtains execution information is a general foreseeable time point. The connection between the observation node and the priori foreseeable time point (i.e., the end time of the priori foreseeable foreseeable action) is a general foreseeable link, and the length of this general foreseeable link is a fixed value o. - That is, the time range in which the observed node can occur is a fixed value Eo. - The interval [o] in the figure - o - The symbol represents the time range within which fixed advance execution information is obtained, i.e., the time range within which the observation node occurs. It indicates that it is known a priori that a random time point will occur after the observation node. - It occurs within a certain time frame, meaning the fixed advance time is 0. - .

[0071] Variable-ahead information refers to the ability of operatives to know the exact duration of an action within a certain timeframe before its known end. For example, when a patrol vehicle is assembling at a target location after receiving a reconnaissance mission, the mission personnel can estimate the exact assembly time 30-50 minutes before reaching the destination based on traffic conditions along the way.

[0072] In the PKSTNU (Primitive Predictable Temporal Network), the time point at which a priori predictable accidental action obtains execution information (i.e., the observation node) corresponding to variable advance execution information is a general accidental time point. The connection between the observation node and the a priori predictable accidental time point (i.e., the time point at which the a priori predictable accidental action ends) is a general accidental link, and the interval length of this general accidental link varies within the range [o...]. - o + The interval [o] in the figure - o +To obtain the time range of variable advance execution information, it is indicated that the a priori known accidental time point will occur after the observation node. - o + It will happen within a certain time frame.

[0073] Controlled-delay information refers to the duration of an action that an operator can obtain within a certain time frame after the start of an accidental action, which is known a priori. The point in time when this information is obtained is the controlled time point. For example, in a mass evacuation mission, the commander can contact the personnel carrying out the mission within 10 minutes after the start of the action and estimate the remaining evacuation time based on the situation on site.

[0074] In the prior-known temporal network PKSTNU, the time point at which the prior-known accidental action obtains execution information (i.e., the observation node) corresponding to the controlled delayed execution information is the controlled time point. The time point between the observation node and the start time point of the prior-known accidental action is a controlled link, and the interval length of this controlled link varies within the range of [o - o + The interval [o] in the figure - o + To obtain the time range of controlled delayed execution information, it means that the observation node will occur after the a priori known start time of the accidental action. - o + It occurs within a certain time period, and the observation node is a controlled time point.

[0075] Contingent-delay refers to the duration of an action that an actor can obtain within a certain time frame after the start of a contingent action, which is known a priori. The time point at which the execution information is obtained is determined by the external environment and is not under control, thus the time point at which the execution information is obtained is contingent. For example, when the target task executors are performing a key location monitoring task, the commander may notify the target task executors of the remaining monitoring time about 30 minutes after the start of the task.

[0076] In the prior-known temporal network PKSTNU, the time point at which the prior-known accidental action obtains execution information (i.e., the observation node) corresponding to the accidental delayed execution information is a general accidental time point. The time point between the observation node and the start time point of the prior-known accidental action is a general accidental link, and the interval length of this general accidental link varies within the range of [o - o + The interval [o] in the figure - o + To obtain the time range of accidental delayed execution information, it means that the observation node will occur after the a priori known start time of the accidental action. - o +It occurs within a certain time period, and the observation point is a randomly controlled time point.

[0077] Specifically, this invention employs the PKDC-Check algorithm to add constraints that the execution information must satisfy to the prior known temporal network PKSTNU based on the category of the obtained execution information and the execution information propagation constraints. For example... Figures 2(A)-2(D) As shown, the PKDC-Check algorithm of the present invention includes:

[0078] For prior known accidental actions in the resource scheduling scheme where fixed advance execution information is available, delete the prior known accidental link between the start and end points of the prior known accidental action, and add a general accidental link from the start point of the prior known accidental action to the observation node. The interval of the added general accidental link is [ao]. - ,bo - As shown in Figure 2(A), in the fixed-ahead case, the prior known accidental link between XE in the fixed-ahead case is deleted, and a general accidental link is added between XO. Since the prior known accidental time point E is after the observation node O, - The time occurs, at which point O must be after point X [ao] - ,bo - It occurs within the range of [XO]. Since the occurrence time of point O is determined by the environment, the time interval between XO is controlled by the environment, and XO are connected by general accidental links.

[0079] The difference between a priori known accidental actions in resource scheduling schemes where variable advance execution information is available and those where fixed advance execution information is available lies in the addition of a general accidental link interval [ao]. - ,bo + Simultaneously, it is necessary to change the general accidental links between observation nodes and a priori known accidental time points (i.e., the end time of a priori known accidental actions) to priori known accidental links, and add corresponding observation nodes to these priori known accidental links. As shown in Figure 2(B), in the variant-ahead case, delete the priori known accidental links between XE, add a general accidental link between XO, and modify the general accidental links between OE to priori known accidental links, adding a new observation node O'. Since the priori known accidental time point E is after observation node O, [o - o + The time occurs, at which point O must be after point X. - ,bo +The event occurs within the specified range. Since the occurrence time of point O is determined by the environment, the time interval between XO is controlled by the environment, and XO is connected using a general accidental link. Since the occurrence time of point E can be determined after obtaining the exact execution information at point O (i.e., determining the end time of the a priori known accidental action), the link between O and E is modified to a priori known accidental link.

[0080] For a priori known accidental actions in a resource scheduling scheme where controlled delay execution information is available, there exists at least one controlled time point. This controlled time point occurs before the priori known accidental time point, and its latest occurrence time is after the observation node. An observation-controlled link is added between the observation node and the controlled time point that needs to occur before the priori known accidental time point (i.e., the time when the priori known accidental action ends). The length of this observation-controlled link varies within the range [min(bo]]. — -v,max(ro — -v,0)),ro — -u], simultaneously, delete the link between a priori known accidental time points and controlled time points that need to occur before the a priori known accidental time points, modify the a priori known accidental link between the a priori known accidental action start time point and the a priori known accidental time point to a general accidental link; change the controlled link interval between the observation node and the a priori known accidental action start time point to [o - o - ], that is, changed to a fixed value o - In this context, the observed controlled link is a constraint that must be satisfied at a controlled time point that occurs before a priori known accidental time point. As shown in Figure 2(C), in the controlled-dealy case, the interval of the XO link is modified to [o - o - A controlled observation link is added between O and Y, and the link between Y and E is deleted. The prior known accidental link between X and E is modified to a general accidental link. Since the observation node O is a controlled time point, the execution information obtained at point O helps to make the network dynamically controllable. Therefore, the network is dynamically controllable if and only if point O becomes dynamically controllable at the earliest time. In this case, the occurrence time of point Y needs to be decided based on the execution information observed at point O.

[0081] The difference between prior knowledge of accidental actions in resource scheduling schemes, where accidental delay information is available, and obtaining controlled delay information, lies in the fact that the observed range of controlled link length variation is [min(bo...]. + -v,max(ro + -v,0)),ro + -u] modifies the interval of the general accidental link between the observed node and the a priori known start time of the accidental action to [o].+ o + ], that is, modified to a fixed value o + As shown in Figure 2(D), in the contingent-delay case, the XO link interval changes from the original [o] - o + ] changed to [o + o + A controlled observation link is added between O and Y, and the link between Y and E is deleted. The prior known accidental link between X and E is modified to a general accidental link. Since the observation node O is an accidental time point, the execution information obtained at point O helps to make the network dynamically controllable. Therefore, the network is dynamically controllable if and only if point O becomes dynamically controllable at the latest time. In this case, the occurrence time of point Y needs to be decided based on the execution information observed at point O.

[0082] This invention employs the DC-Check algorithm to determine the dynamic controllability of a temporal network containing execution information. The DC-Check algorithm includes:

[0083] For each type of execution information, it is determined whether the updated temporal network is dynamically controllable when the execution information *r* obtained during the prior known accidental action (i.e., when the prior known duration of the accidental action is *a* and *b*) is both known. If so, the resource scheduling scheme is effective; otherwise, the resource scheduling scheme is ineffective. The dynamic controllability detection algorithm of this invention ensures that the network is dynamically controllable under any execution information by considering the dynamic controllability of all execution information in extreme cases.

[0084] Based on the aforementioned execution information propagation constraints, the prior-knowledgeable temporal network PKSTN can be made to no longer contain prior-knowledgeable observation links and observation links, and it is guaranteed that the value range of all time points in the network does not change. Through the DC-Check algorithm, the observation-controlled links in the network are converted into controlled links (the parameter r in the observation-controlled links is converted into the definite values ​​a and b). That is, after the propagation constraints, the final temporal network does not contain prior-knowledgeable observation links, observation links, and observation-controlled links. Existing dynamic controllability methods can be used to quickly determine whether the final network is dynamically controllable.

[0085] In the propagation constraints, for fixed early execution information and variable early execution information, the a priori known observation links are directly deleted; for controlled delayed execution information and accidental delayed execution information, the a priori known observation links are transformed into general accidental links; and the observation links in all cases are directly deleted.

[0086] This invention also provides a resource scheduling scheme validity verification system based on execution information, mainly comprising:

[0087] The a priori-knowable temporal network construction module is used to construct a priori-knowable temporal network. The a priori-knowable temporal network includes controlled time points, general accidental time points, a priori-knowable accidental time points, controlled links, general accidental links, a priori-knowable accidental links, and observation links. Among them, controlled links point to one of the controlled time points, general accidental time points, or a priori-knowable accidental time points; general accidental links point to general accidental time points or a priori-knowable accidental time points; a priori-knowable accidental links point to a priori-knowable accidental time points; and observation links contain observation nodes, which point to a priori-knowable accidental links.

[0088] The resource scheduling scheme temporal constraint module is used to define the start time of the execution action in the resource scheduling scheme and the end time of the action with a definite duration as controlled time points. Actions with uncertain durations that can obtain execution information are categorized as a priori known accidental actions, while those that cannot obtain execution information are categorized as accidental actions. The end time of a priori known accidental action is defined as a priori known accidental time point, the end time of an accidental action is defined as a general accidental time point, and the time point at which execution information is obtained is defined as an observation node. The start and end points of actions with definite durations are connected by controlled links, the start and end points of accidental actions are connected by general accidental links, the start and end points of priori known accidental actions are connected by priori known accidental links, and any two actions are connected by controlled links.

[0089] The network constraint propagation module is used to add constraints that the prior-known temporal network must satisfy based on the category of the obtained execution information, in order to update the duration of prior-known accidental actions. Here, the execution information refers to information learned during the execution of actions with uncertain durations that can reduce the uncertainty of the action's duration.

[0090] The dynamic controllability detection module is used to determine the dynamic controllability of the updated temporal network. If the updated temporal network is controllable, the resource scheduling scheme is effective; otherwise, the resource scheduling scheme is invalid.

[0091] The network constraint propagation module includes one or more of the following: a variable early execution information constraint propagation unit, a variable early execution information propagation unit, a controlled delayed execution information propagation unit, and an accidental delayed execution information propagation unit;

[0092] A variable advance execution information constraint propagation unit is used to delete a priori known accidental links between the a priori known accidental action's start and end points; and to add general accidental links from the a priori known accidental action's start point to the observation node, where the interval of the general accidental link is [ao]. - ,bo - ];

[0093] A variable advance execution information propagation unit is used to delete a priori known accidental links between the a priori known start and end points of accidental actions; and to add general accidental links from the start point of a priori known accidental action to the observation node, wherein the interval of the general accidental links is [ao]. - ,bo + ]; Change the general accidental link from the observation node to the a priori known accidental time point to the a priori known accidental link, and add the observation node corresponding to the a priori known accidental link;

[0094] The controlled delayed execution information propagation unit is used to add an observation-controlled link between the observation node and a controlled time point that occurs before a priori known accidental time point; wherein, the controlled time point must also satisfy the constraint that it occurs no later than the observation node, and the observation-controlled link is a constraint that the controlled time point must satisfy; delete the link between the priori known accidental time point and the controlled time point; modify the priori known accidental link between the priori known accidental action start time point and the priori known accidental time point to a general accidental link; and set the controlled link interval between the observation node and the priori known accidental action start time point to a fixed value o. - ;

[0095] The accidental delayed execution information propagation unit is used to add an observation-controlled link between the observation node and a controlled time point that occurs before a priori known accidental time point; wherein, the controlled time point must also satisfy the constraint that it occurs no later than the observation node, and the observation-controlled link is a constraint that the controlled time point must satisfy; delete the link between the priori known accidental time point and the controlled time point; modify the priori known accidental link between the priori known accidental action start time point and the priori known accidental time point to a general accidental link; and set the controlled link interval between the observation node and the priori known accidental action start time point to a fixed value o. + ;

[0096] Among them, o - This represents the minimum value of the difference between the observed node and the time point connected to it, o. + denoted by , represents the maximum value of the difference between the observation node and the time point connected to the observation node, and 'a' and 'b' represent the minimum and maximum values ​​of the range of values ​​for the duration of a priori known accidental action, respectively.

[0097] The method for verifying the effectiveness of a resource scheduling scheme based on execution information, as described in this invention, generally includes a temporal network construction process for the resource scheduling scheme and a dynamic controllability detection process. In the temporal network construction process, a novel prior-knowable temporal network (PKSTN) is constructed. This PKSTN includes three time points and four links. Prior-knowable accidental links and observation links represent execution actions with uncertain durations and observation processes where execution information can be obtained during execution. Through propagation constraints, the temporal constraints that need to be met to address temporal uncertainty using execution information and all controlled time points are derived. A dynamic controllability detection algorithm verifies whether the resource scheduling scheme meets the conditions for dynamic controllability when facing temporal uncertainty, thereby determining whether the resource scheduling scheme can be successfully executed after reducing uncertainty by utilizing the execution information obtained during execution. Because the method of this invention considers the execution information that may be obtained during execution in the verification process, when applied to actual resource scheduling schemes to verify their effectiveness, almost no abnormal situations or errors occur, improving the accuracy and efficiency of the verification.

[0098] Since the method of the present invention can automatically verify the effectiveness of resource scheduling schemes by computer, it has a low degree of dependence on human intervention and informal knowledge, and therefore has long-term stability.

[0099] Furthermore, the method of the present invention is suitable for verifying the effectiveness of resource scheduling schemes under large-scale tasks.

[0100] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for validating the effectiveness of a resource scheduling scheme based on execution information, characterized in that, include: A priori-knowable temporal network is constructed, which includes controlled time points, general accidental time points, priori-knowable accidental time points, controlled links, general accidental links, priori-knowable accidental links, and observation links. Among them, controlled links point to one of the controlled time points, general accidental time points, or priori-knowable accidental time points; general accidental links point to either general accidental time points or priori-knowable accidental time points; priori-knowable accidental links point to priori-knowable accidental time points; and observation links contain observation nodes, which point to priori-knowable accidental links. The start time of the target task execution action in the resource scheduling scheme and the end time of the action with a definite duration are taken as the controlled time point; the action with an uncertain duration that can obtain execution information is recorded as a priori known accidental action, and the action that cannot obtain execution information is recorded as accidental action. The end time of the priori known accidental action is taken as the priori known accidental time point, the end time of the accidental action is taken as the general accidental time point, and the time point of obtaining execution information is taken as the observation node. The start and end points of actions with a definite duration are connected by controlled links; the start and end points of accidental actions are connected by general accidental links; the start and end points of a priori accidental actions are connected by a priori accidental links; and any two actions are connected by controlled links. Based on the category of the obtained execution information, constraints that the execution information must satisfy are added to the prior known temporal network to update the duration of the prior known accidental action, and the dynamic controllability of the updated temporal network is judged. If the temporal network is controllable, the resource scheduling scheme is effective; otherwise, the resource scheduling scheme is invalid. The execution information is information obtained during the execution of an action with an uncertain duration that can reduce the uncertainty of the duration of the action.

2. The method according to claim 1, characterized in that, The categories of execution information include any one of: fixed early execution information, variable early execution information, controlled delayed execution information, and occasional delayed execution information; The fixed advance execution information refers to the fact that the personnel executing the target task know the end time of the prior known accidental action at a fixed time before the end of the accidental action; The variable advance execution information refers to the duration of the prior known accidental action that the target task executor knows within a certain time range before the end of the prior known accidental action; The controlled delayed execution information refers to the duration of the prior known accidental action that the target task executor learns within a certain time range after the start of the prior known accidental action, and the time point at which the execution information is learned is a controlled time point; The term "accidental delayed execution information" refers to the information obtained by the target task executor within a certain time range after the prior known accidental action has begun, indicating that the duration of the prior known accidental action is accidental, and that the time point at which the execution information is obtained is an accidental time point.

3. The method according to claim 2, characterized in that, In the prior known temporal network, When variable advance execution information is obtained, the observation node is a general accidental time point, and there is a general accidental link between the observation node and the prior known accidental time point, and the length of the general accidental link is a fixed value o. - ; When variable advance execution information is obtained, the observation node is a general accidental time point, and there is a general accidental link between the observation node and the prior known accidental time point, and the length of the general accidental link varies within the range of [o]. - o + ]; When controlled delay execution information is obtained, the observation node is a controlled time point, and the observation node and the time point at which the prior known accidental action begins are connected by a controlled link, with the length of the controlled link varying within the range of [o]. - o + ]; When accidental delayed execution information is obtained, the observation node is a general accidental time point, and the observation node and the time point at which the prior known accidental action begins are general accidental links, with the length of the general accidental link varying within the range of [o]. - o + ]; Among them, o - This represents the minimum value of the difference between the observed node and the time point connected to it, o. + This represents the maximum value of the difference between the observed node and the time point connected to it.

4. The method according to claim 3, characterized in that, When variable advance execution information is obtained, the constraints added to the prior known temporal network are as follows: Remove the prior known accidental link between the prior known start and end points of accidental actions; Add a general accidental link from the starting point of the accidental action, which is known a priori to the observation node, to the observation node, and the length of the general accidental link varies within the range of [ao]. - ,bo - ], where a and b represent the minimum and maximum values ​​of the range of values ​​for the duration of a known accidental action, respectively.

5. The method according to claim 3, characterized in that, When variable advance execution information is obtained, the constraints added to the prior known temporal network are as follows: Remove the prior known accidental link between the prior known start and end points of accidental actions; Add a general accidental link from the starting point of the accidental action, which is known a priori to the observation node, to the observation node, and the length of the general accidental link varies within the range of [ao]. - ,bo + ], where a and b represent the minimum and maximum values ​​of the range of values ​​for the duration of a known accidental action, respectively; Change the general accidental link from the observation node to the a priori known accidental time point to a priori known accidental link, and add the observation node corresponding to the a priori known accidental link.

6. The method according to claim 3, characterized in that, When controlled delay execution information is obtained, the constraints added to the prior known temporal network are as follows: There exists at least one controlled time point, which occurs before a priori known accidental time point, and the controlled time point occurs at the latest after the observation node. An observation-controlled link is added between the observation node and the controlled time point; wherein, the observation-controlled link is a constraint that the controlled time point needs to satisfy. Delete the link between the previously known accidental time point and the controlled time point; The prior known accidental link between the known start time of the accidental action and the known accidental time point is modified to a general accidental link; The controlled link length variation range between the observation node and the prior known start time of the accidental action is changed to a fixed value. - .

7. The method according to claim 3, characterized in that, When information about accidental execution delays is obtained, the constraints added to the prior known temporal network are as follows: There exists at least one controlled time point, which occurs before a priori known accidental time point, and the controlled time point occurs at the latest after the observation node. An observation-controlled link is added between the observation node and the controlled time point; wherein, the observation-controlled link is a constraint that the controlled time point needs to satisfy. Delete the link between the previously known accidental time point and the controlled time point; The prior known accidental link between the known start time of the accidental action and the known accidental time point is modified to a general accidental link; The controlled link length variation range between the observation node and the prior known start time of the accidental action is changed to a fixed value. + .

8. The method according to claim 1, characterized in that, Determining the dynamic controllability of the updated temporal network includes: If the duration of the prior known accidental action is a and b respectively, it is determined whether the prior known temporal network is dynamically controllable. If so, the resource scheduling scheme is effective; otherwise, the resource scheduling scheme is invalid. Here, a and b represent the minimum and maximum values ​​of the range of the duration of the prior known accidental action, respectively. Wherein, the execution information obtained by the prior known accidental action is any one of fixed early execution information, variable early execution information, controlled delayed execution information, and accidental delayed execution information.

9. A system for verifying the effectiveness of a resource scheduling scheme based on execution information, characterized in that, include: A priori-knowable temporal network construction module is used to construct a priori-knowable temporal network, which includes controlled time points, general accidental time points, priori-knowable accidental time points, controlled links, general accidental links, priori-knowable accidental links, and observation links. Among them, controlled links point to controlled time points, general accidental time points, or priori-knowable accidental time points; general accidental links point to general accidental time points or priori-knowable accidental time points; priori-knowable accidental links point to priori-knowable accidental time points; and observation links contain observation nodes, which point to priori-knowable accidental links. The temporal constraint module of the resource scheduling scheme is used to define the start time of the target task execution action and the end time of the action with a definite duration as controlled time points in the resource scheduling scheme. Actions with uncertain durations that can obtain execution information are categorized as a priori known accidental actions, while those that cannot obtain execution information are categorized as accidental actions. The end time of a priori known accidental action is defined as a priori known accidental time point, the end time of an accidental action is defined as a general accidental time point, and the time point of obtaining execution information is defined as an observation node. The start and end points of actions with definite durations are connected by controlled links, the start and end points of accidental actions are connected by general accidental links, the start and end points of priori known accidental actions are connected by priori known accidental links, and any two actions are connected by controlled links. The network constraint propagation module is used to add constraints that the obtained execution information must satisfy to the prior known temporal network, based on the category of the execution information, in order to update the duration of the prior known accidental actions. The execution information refers to information learned during the execution of actions with uncertain durations that can reduce the uncertainty of the action's duration. The dynamic controllability detection module is used to determine the dynamic controllability of the updated temporal network. If the updated temporal network is dynamically controllable, the resource scheduling scheme is effective; otherwise, the resource scheduling scheme is invalid.

10. The system according to claim 9, characterized in that, The network constraint propagation module includes one or more of the following: a fixed early execution information constraint propagation unit, a variable early execution information propagation unit, a controlled delayed execution information propagation unit, and an occasional delayed execution information propagation unit; The variable advance execution information constraint propagation unit is used to delete the a priori known accidental link between the a priori known accidental action's start and end points; and to add a general accidental link from the a priori known accidental action's start point to the observation node, wherein the length of the general accidental link varies within the range of [ao]. - ,bo - ]; The variable advance execution information propagation unit is used to delete the a priori known accidental link between the a priori known accidental action's start and end points; and to add a general accidental link from the a priori known accidental action's start point to the observation node, wherein the length of the general accidental link varies within the range of [ao]. - ,bo + ]; Change the general accidental link from the observation node to the a priori known accidental time point to the a priori known accidental link, and add the observation node corresponding to the a priori known accidental link; The controlled delayed execution information propagation unit is used to add an observation-controlled link between the observation node and a controlled time point that occurs before a priori known accidental time point; wherein, the controlled time point also satisfies the condition that it occurs no later than the observation node, and the observation-controlled link is a constraint that the controlled time point must satisfy; delete the link between the priori known accidental time point and the controlled time point; modify the priori known accidental link between the priori known accidental action start time point and the priori known accidental time point to a general accidental link; and set the length of the controlled link between the observation node and the priori known accidental action start time point to a fixed value o. - ; The accidental delayed execution information propagation unit is used to add an observation-controlled link between the observation node and a controlled time point that occurs before a priori known accidental time point; wherein, the controlled time point also satisfies the condition that it occurs no later than the observation node, and the observation-controlled link is a constraint that the controlled time point must satisfy; delete the link between the priori known accidental time point and the controlled time point; modify the priori known accidental link between the priori known accidental action start time point and the priori known accidental time point to a general accidental link; and set the length of the controlled link between the observation node and the priori known accidental action start time point to a fixed value o. + ; Among them, o - This represents the minimum value of the difference between the observed node and the time point connected to it, o. + denoted by , represents the maximum value of the difference between the observation node and the time point connected to the observation node, and 'a' and 'b' represent the minimum and maximum values ​​of the range of values ​​for the duration of a priori known accidental action, respectively.

Images