Message increment planning method

Abstract

The invention discloses a message increment planning method, which comprises the following steps of: S1, decomposing a message into Smessages according to all links through which a corresponding virtual link passes, and obtaining a windowing size and a minimum frame interval corresponding to the Smessages; s2, dividing all the messages in the network according to a preset division rule, judging whether all the messages are divided into a specified set or not, and if so, executing a step S3; s3, an SMT solver is constructed, and initial auxiliary constraint of the SMT solver is carried out; s4, corresponding constraint conditions are set according to the corresponding sets, and a set-constraint SMT solving model is generated; and S5, inputting the data of the corresponding set into the set-constraint SMT solving model obtained in the step S4 by applying the set-constraint SMT solving model, and calculating to obtain an information planning result. According to the invention, through message division, constraint design and fractional planning, the message planning time can be shortened, and the network message service scale which can be actually planned is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of message planning methods, and in particular to a message increment planning method. Background Art

[0002] Time-Triggered Ethernet (TTEthernet) is a communication network technology developed based on traditional Ethernet. It aims to provide an environment for the coexistence of time-triggered and event-triggered services by establishing a high-precision global clock synchronization mechanism, and ensure the time determinism of real-time services.

[0003] Specifically, TTEthernet first introduces the high-precision clock synchronization service provided by the SAE AS6802 standard or the high-precision clock synchronization mechanism of the PTP (Precision Time Protocol) specified in the IEEE1588 standard to the non-deterministic traditional Ethernet, making ordinary Ethernet an underlying network with a globally unified time base. It then adds time-triggered real-time services (TT) and event-triggered non-real-time services (ET) for transmission services, and adopts a reasonable scheduling strategy to ensure that message transmission is conflict-free and mutually exclusive. In other words, it ensures that time-triggered services are scheduled according to a predetermined schedule, ensuring their determinism in time and sequence, while allowing event-triggered services to be transmitted dynamically based on the events that occur.

[0004] The current mainstream TTE scheduling generation method is based on a time schedule generation method using an SMT (Satisfiability Modulo Theories) solver. This method, proposed by Wilfried Steiner of TTTech, addresses the scheduling problem of time-triggered (TT) services in Time-Triggered Ethernet (TTethernet). This method validates the validity of the planned time schedule by setting constraints and optimizing the objective function. The implementation process abstractly describes the service parameters of the message to be scheduled in the network, the network topology, and the scheduling time constraints at each communication node, forming the input parameters for the SMT solver. Based on these input parameters, the SMT solver uses its internal decision-making process to calculate whether the message can successfully allocate bandwidth resources in the time-triggered network, thereby determining the satisfiability of the scheduling problem. If the judgment is true, it then generates specific scheduling time points, and this process continues until the entire time schedule is constructed.

[0005] However, a major drawback of this approach is that the SMT solver uses an exhaustive search in the state space to perform calculations. As the network and service scale increases, the computational time complexity increases dramatically, resulting in low algorithm efficiency and even the inability to find a solution. Therefore, although the SMT solver-based approach provides an effective theoretical basis and technical means for scheduling TT services, in practical applications, especially when facing large-scale networks, its computational performance and solution capabilities may be limited. Currently, scholars have developed a large-scale solution method for SMT problems based on differential evolution. However, this method loses precision and introduces manual intervention, making it unsuitable for TTE planning problems, which require strict precision in planning results. Summary of the Invention

[0006] The present invention aims to at least partially address one of the technical problems in the related art. To this end, one objective of the present invention is to propose a method for incremental message planning. By partitioning messages, designing constraints, and implementing a phased planning process, this method can shorten message planning time and increase the scale of network message services that can be actually planned.

[0007] According to a message increment planning method proposed by the present invention, the method steps are as follows:

[0008] S1: Decompose the message Message into Smessage according to all the links that the corresponding virtual link passes through, and obtain the corresponding window size and minimum frame interval of Smessage;

[0009] S2: Apply Smessage obtained in step S1 to divide all messages in the network according to the preset division rules, and determine whether all messages are divided into the specified set. If so, execute step S3;

[0010] S3: Build the SMT solver, set the initial scope of the SMT solver, and perform the initial auxiliary constraints of the SMT solver;

[0011] S4: Apply the set data obtained in step S2 and the initial auxiliary constraint SMT solver obtained in step S3, set corresponding constraints according to the corresponding sets, and generate a set-constraint SMT solution model;

[0012] S5: Apply the set-constraint SMT solution model obtained in step S4, input the data of the corresponding set into the set-constraint SMT solution model to calculate and obtain the information planning result.

[0013] Preferably, in step S1:

[0014] S11: Obtain the message transmission path according to the virtual link VL of the message;

[0015] S12: Apply the message transmission path obtained in step S11, and generate a corresponding Smessage for each physical link passed through the message transmission path;

[0016] S13: Apply the Smessage obtained in step S12 and calculate the corresponding window size and minimum frame interval according to the Smessage transmission rate and message attributes.

[0017] Preferably, in step S2:

[0018] S21: Select messages with mandatory constraints on the sending timing from all TT messages, and decompose the selected messages to generate all Smessage sets as the first set vectorGNC of message planning;

[0019] S22: Apply the process of filtering the remaining TT messages in step S21, analyze the remaining TT message loads, select the device port with the largest message load, and select all Smessage sets generated by decomposing the messages corresponding to all virtual links (VLs) passing through this link as the second set (vectorCongestion) of message planning;

[0020] S23: Apply all the Smessage sets generated by decomposing the remaining messages after screening in step S21 and step S22 as the third set vectorOther of message planning.

[0021] Preferably, in step S22:

[0022] TTE network uses full-duplex network. For a message sending and receiving port,<devid,portid> To uniquely identify each Smessage, each Smessage also corresponds to a unique message sending and receiving port. Statistics are performed on the Smessages corresponding to the remaining TT messages filtered in step S21. The formula for calculating the load of each link is as follows:

[0023]

[0024] Where L represents the load of the link, window i Indicates the window time of the i-th Smessage on the current link, TTGap i Indicates the frame interval of the Smessage, clusterCycle is the cluster period of the TTE network, period i The message cycle corresponding to the Smessage.

[0025] Preferably, in step S4, the constraints include: single cycle constraint, offset constraint, conflict-free constraint, message window constraint, timing requirement constraint and increment constraint;

[0026] Define S as the set of Smessage variables for planning, sm is one of the Smessage variables, dispatch_point is the sending time point, window is the time window size, TTGap is the message interval, clusterCycle is the cluster period of the TTE network, period i is the message cycle corresponding to the i-th Smessage;

[0027] The single-cycle constraint is:

[0028]

[0029] The offset constraints are:

[0030] The period of TT message is an integer multiple of the network integration period inteCycle:

[0031] period=n×inteCycle,n∈Z,n>0

[0032] For each message there should be:

[0033]

[0034] Where period represents the message cycle, n is an integer indicating that the message cycle is a multiple of the integration cycle, inteCycle represents the integration cycle of network messages, k represents the kth integration cycle, shift_time represents the time period at the beginning of each integration cycle that cannot be used for message scheduling, and dispatch_point represents the time when the message is sent.

[0035] The conflict-free constraints are:

[0036] (dispatch_point i +window i +TTGap i +k×period i

[0037] ≤dispatch_point j +l×period j )

[0038] OR(dispatch_point j +window j +TTGap j +l×periodj

[0039] ≤dispatch_point i +k×period i )

[0040] Here, k represents the kth sending of sub-message i, and l represents the lth sending of sub-message j.

[0041] The message window constraint is: set according to the message forwarding mechanism of the actual network device;

[0042] Timing requirement constraints are: set according to actual message service requirements;

[0043] The increment constraints are:

[0044]

[0045] Among them, pre_result i Indicates sm i Pre-planning results.

[0046] Preferably, in step S4, the set-constraint SMT solution model includes: using single-cycle constraints, offset constraints, conflict-free constraints, message window constraints and timing requirement constraints for the data in the first set vectorGNC; using single-cycle constraints, offset constraints, conflict-free constraints, message window constraints and incremental constraints for the data in the second set vectorCongestion; using single-cycle constraints, offset constraints, conflict-free constraints, message window constraints and incremental constraints for the data in the third set vectorOther.

[0047] Preferably, in step S3:

[0048] Initial auxiliary constraints:

[0049]

[0050] The beneficial effects of the present invention are:

[0051] (1) By dividing messages, designing constraints, and implementing a phased planning process, we can shorten message planning time and increase the scale of network message services that can actually be planned.

[0052] (2) Message division and selection are automatically generated by link load calculation, without the need for manual intervention;

[0053] (3) Compared with the traditional SMT-based TTE message planning method that cannot directly obtain the solver status, since the messages are divided into three groups and planned in stages, the message planning progress can be observed by observing the result files generated by each planning and the software output. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In the attached figure:

[0055] Figure 1 This is a flowchart of a message incremental planning method proposed by the present invention;

[0056] Figure 2 This is an example diagram of decomposing the message Message proposed by the present invention into Smessage;

[0057] Figure 3 This is an example diagram of Smessage1 and Smessage2 under the conflict-free constraint proposed by the present invention;

[0058] Figure 4 This is an example diagram of dividing all messages in the network according to rules proposed by the present invention;

[0059] Figure 5 The present invention proposes Figure 4 Example diagram of message constraints under the network conditions shown;

[0060] Figure 6 This is a comparison chart of the solution time of the step-by-step planning proposed by the present invention and the existing direct planning solver as the number of messages changes. DETAILED DESCRIPTION

[0061] Reference Figure 1 , a message increment planning method, the method steps are as follows:

[0062] S1: Decompose the message Message into Smessage according to all the links that the corresponding virtual link passes through, and obtain the corresponding window size and minimum frame interval of Smessage;

[0063] In this embodiment:

[0064] S11: Obtain the message transmission path according to the virtual link VL of the message;

[0065] S12: Apply the message transmission path obtained in step S11, and generate a corresponding Smessage for each physical link passed through the message transmission path;

[0066] S13: Apply the Smessage obtained in step S12 and calculate the corresponding window size and minimum frame interval according to the Smessage transmission rate and message attributes.

[0067] Specifically, such as Figure 2 As shown, the multicast message VL sent from ES_1 to ES_4 and ES_5 along the illustrated path can be decomposed into five Smessages according to the links.

[0068] S2: Apply Smessage obtained in step S1 to divide all messages in the network according to the preset division rules, and determine whether all messages are divided into the specified set. If so, execute step S3;

[0069] In this embodiment:

[0070] S21: Select messages with mandatory constraints on the sending timing from all TT messages, such as GNC service messages, and decompose the selected messages to generate all Smessage sets as the first set vectorGNC of message planning;

[0071] S22: Apply the process of filtering the remaining TT messages in step S21, analyze the remaining TT message loads, select the device port with the largest message load, and select all Smessage sets generated by decomposing the messages corresponding to all virtual links (VLs) passing through this link as the second set (vectorCongestion) of message planning;

[0072] Specifically:

[0073] TTE network uses full-duplex network. For a message sending and receiving port, you can use<devid,portid> Each Smessage is uniquely identified and corresponds to a unique message sending and receiving port. The formula for calculating the load of each link by counting all Smessages is as follows:

[0074]

[0075] Where L represents the load of the link, window i Indicates the window time of the i-th Smessage on the current link, TTGap i Indicates the frame interval of the Smessage, clusterCycle is the cluster period of the TTE network, period i The message cycle corresponding to the Smessage.

[0076] In the formula (window i +TTGap i ) indicates the length of time that a message is sent on the sending and receiving port, clusterCycle / period i It represents the total number of times a message is sent in a cluster cycle. The product of the two gives the total time a message occupies on a transmitting and receiving port in a cluster cycle, that is, its link load.

[0077] S23: Apply all the Smessage sets generated by decomposing the remaining messages after screening in step S21 and step S22 as the third set vectorOther of message planning.

[0078] From this, you can evaluate and find the most congested device ports.

[0079] More specific:

[0080] like Figure 4 As shown, if the link from SW_1 to SW_5 in the network is calculated<SW_1,SW_5> If the load is the largest, then all Smessages decomposed from the VL containing this link will be filled into vectorCongestion.

[0081] S3: Build the SMT solver, set the initial scope of the SMT solver, and perform the initial auxiliary constraints of the SMT solver;

[0082] In this embodiment:

[0083] During the actual search process, an SMT solver assigns values to variables and continuously checks for constraint conflicts. Without an initial scope, the solver may waste significant time in areas unrelated to the actual constraints. For TTE network message planning, a simplified offset constraint is used to constrain the initial values, enabling the solver to find a solution more efficiently.

[0084] More specific:

[0085] The initial auxiliary constraints are:

[0086]

[0087] S4: Apply the set data obtained in step S2 and the initial auxiliary constraint SMT solver obtained in step S3, set corresponding constraints according to the corresponding sets, and generate a set-constraint SMT solution model;

[0088] In this embodiment:

[0089] Constraints include: single cycle constraints, offset constraints, conflict-free constraints, message window constraints, timing requirement constraints, and increment constraints;

[0090] Planning the variables fed into the SMT solver requires setting appropriate constraints to meet the requirements of TTE network message transmission. Let S be the set of Smessage variables to be planned, sm is one of the Smessage variables, dispatch_point is the sending time, window is the time window size, and TTGap is the message interval. It's important to note that within a clusterCycle period, a message is not sent just once, but clusterCycle / period times. The dispatch_point here refers to the first sending point within the message period.

[0091] The single-cycle constraint is:

[0092] The purpose of the single-cycle constraint is to limit the sending of messages to be completed within one message cycle to meet the nature of periodic sending of TT messages, which complies with the following requirements:

[0093]

[0094] The offset constraints are:

[0095] In a TTE network, messages cannot be sent at any time. At the beginning (or end, depending on the actual hardware implementation) of each integration cycle, there is a period of time used to execute the clock synchronization protocol. The time window for TT message planning cannot fall within this period.

[0096] The period of TT message is an integer multiple of the network integration period inteCycle:

[0097] inteCycle=n×period,n∈Z,n>0

[0098] For each message there should be:

[0099]

[0100] Where period represents the message cycle, n is an integer indicating that the message cycle is a multiple of the integration cycle, inteCycle represents the integration cycle of network messages, k represents the kth integration cycle, shift_time represents the time period at the beginning of each integration cycle that cannot be used for message scheduling, and dispatch_point represents the time when the message is sent.

[0101] The conflict-free constraints are:

[0102] Conflict-free constraints are the most numerous and computationally intensive part of message planning. Their purpose is to ensure that the time windows set for TT messages on the sending ports of network devices are exclusive and conflict-free. When setting these constraints, each message instance sent throughout the entire cluster cycle is compared. The number of constraints passed to the solver increases at the square of the number of message instances compared. This method uses interval comparison to filter the comparison of message instances, removing constraints that will not actually cause collisions and reducing the amount of computational effort required for comparisons.

[0103] like Figure 3 As shown, Smessage1 and Smessage2 are sent three and two times, respectively, on the same link within a cluster cycle. When performing conflict-free constraints, there's no need to compare each instance of Smessage1 with each instance of Smessage2, requiring C(3,1) × C(2,1) = 3 × 2 = 6 calculations within the solver. Instead, only the overlapping periods need to be compared: Smessage2_1 is compared with Smessage1_1 and Smessage1_2, and Smessage2_2 is compared with Smessage1_2 and Smessage1_3. This requires only four comparisons within the solver. This process significantly improves the efficiency of the SMT solver by reducing the number of constraint comparisons.

[0104] In traditional methods, the SMT solver needs to consider all message instances and the constraints between them, which will lead to a large amount of redundant calculations. By focusing only on the overlapping parts within the cycle, the solver can avoid meaningless comparisons of irrelevant message instances, thereby reducing the number of constraints and variables and reducing computational complexity. This not only reduces the number of constraints that the solver needs to handle, but also makes the conflict detection and propagation process more efficient, because the solver only needs to focus on those instances where conflicts may occur, avoiding wasting computing resources on non-conflicting constraints. In addition, this optimization also accelerates the solver's pruning and heuristic search, reducing the size of the search space, making the solution process more efficient and able to converge to the solution more quickly. Ultimately, the solver can complete the solution of complex scheduling problems in a shorter time, especially in complex scenarios such as large-scale cluster scheduling. This method effectively improves the solution efficiency.

[0105] For Smessages that need to be compared i The kth sending instance and Smessage j The lth sending instance,

[0106] The following comparison constraint is set. Specifically, the end time of the former (or latter) time window must be TTGap before the start time of the latter (or former) time window. For example, to ensure that the interval between the two time windows is TTGap, in order to meet the requirement of non-conflicting TTE network message sending, the specific expression is as follows:

[0107] (dispatch_point i +window i +TTGap i +k×period i

[0108] ≤dispatch_point j +l×period j )

[0109] OR(dispatch_point j +window j +TTGap j +l×period j

[0110] ≤dispatch_point i +k×period i )

[0111] Where k represents the kth sending of sub-message i, and l represents the lth sending of sub-message j;

[0112] The message window constraint is defined as follows: The transmission window time on each network device is not completely independent of the message propagation across the entire virtual link. Because the basic variables used in planning are separate Smessages derived from the message, constraints are required to connect them in series. The setting of this constraint is related to the message forwarding mechanism of the actual network devices.

[0113] Timing constraints: Special messages, such as GNC, often have specific constraints on their delivery timing, limiting the range of delivery times and the order in which they are delivered. This constraint is related to the actual message service requirements, but due to the high timing requirements, it should be considered in message planning.

[0114] The incremental constraint is: decomposing the message plan into three plans requires using incremental planning to link the messages. Constraints are used to fix the time window of the planned variables.

[0115]

[0116] Among them, pre_result i Indicates smi Pre-planning results.

[0117] S5: Apply the set-constraint SMT solution model obtained in step S4, input the data of the corresponding set into the set-constraint SMT solution model to calculate and obtain the information planning result.

[0118] In order to more clearly illustrate the scheme and effect of this implementation, the following examples are provided with reference to the accompanying drawings:

[0119] For the attached Figure 4 The network shown has the following messages:

[0120]

[0121]

[0122] The message of VL_1 is sent exactly 1ms earlier than the message of VL_7 and 1.5ms earlier than the message of VL_8.

[0123] For this network, message periods are 4ms, 8ms, and 20ms. The integration period is the greatest common factor of these three, 4ms, and the clustering period is the least common multiple of these three, 40ms. For simplification, the total time a message takes on a link is uniformly defined as t. In practice, this time depends on factors such as message length and hardware forwarding mechanisms.

[0124] According to the formula:

[0125]

[0126] The link with the highest link load is<SW_5,SW_1> , the load is:

[0127]

[0128] The Smessages involved are: SM_2, SM_8, SM_12, SM_33, SM_38, and the corresponding VLs are VL_1, VL_2, VL_3, VL_7, and VL_8.

[0129] Based on the above information, the planned Smessages can be divided into three groups of planning objects.

[0130] The first group is the Smessage set corresponding to the message virtual links VL_1, VL_7, and VL_8 with timing requirements: {SM_1, SM_2, SM_3, SM_4, SM_5, SM_32, SM_33, SM_34, SM_35,

[0131] SM_36,SM_37,SM_38,SM_39,SM_40,SM_41,SM_42}

[0132] The constraints set for it include: single cycle constraint, offset constraint, conflict-free constraint, message window constraint and timing requirement constraint.

[0133] The second group is the Smessage set corresponding to the VL involving the congested link:

[0134] {SM_1,SM_2,SM_3,SM_4,SM_5,SM_6,SM_7,SM_8,SM_9,SM_10,SM_11,SM_12,SM_13,SM_32,SM_33,SM_34,SM_35,SM_36,SM_37,

[0135] SM_38,SM_39,SM_40,SM_41,SM_42}

[0136] The constraints set on it are: single cycle constraint, offset constraint, conflict-free constraint, message window constraint and increment constraint.

[0137] The third group is the remaining Smessage collection:

[0138] {SM_14,SM_15,SM_16,SM_17,SM_18,SM_19,SM_20,SM_21,

[0139] SM_22,SM_23,SM_24,SM_25,SM_26,SM_27,SM_28,SM_29,SM_30,

[0140] SM_31}

[0141] The constraints set on it are: single cycle constraint, offset constraint, conflict-free constraint, message window constraint and increment constraint.

[0142] After three planning executions, a feasible solution for network message planning can be obtained.

[0143] The following will Figure 4 The message constraints in the network shown are illustrated as an example, where the unit of the solver planning time value is 1us.

[0144] Single-cycle constraint:

[0145] For SM_1, let the time length (window + TTGap) it occupies be t, then the constraints to be set for it are:

[0146] dispatch_point1+t<4000

[0147] Offset constraint:

[0148] Assume that the time required for clock synchronization in the TTE network is shift_time = 0.2ms. For SM_6, the message period is 20ms, which is five times the integration period of 4ms. The constraints to be set are:

[0149] (200 <dispatch_point6<4000)

[0150] OR(4000+200 <dispatch_point6<8000)

[0151] OR(8000+200 <dispatch_point6<12000)

[0152] OR(12000+200 <dispatch_point6<16000)

[0153] OR(16000+200 <dispatch_point1<20000)

[0154] No conflicting constraints:

[0155] For the same link<SW_5,SW_1> The periods of SM_8 and SM_12 are 20 and 8 respectively. Comparing the possible transmission examples, see the attached figure. Figure 5 shown.

[0156] Assume window + TTGap = t(us). Set the following 6 conflict-free constraints for SM_8 and SM_12:

[0157] Comparison between SM_8_1 and SM_12_1

[0158] (dispatch_point8+t≤dispatch_point 12 )

[0159] OR(dispatch_point 12 +t≤dispatch_point8)

[0160] Comparison between SM_8_1 and SM_12_2

[0161] (dispatch_point8+t≤dispatch_point 12 +8000)

[0162] OR(dispatch_point 12 +8000+t≤dispatch_point8)

[0163] Comparison between SM_8_1 and SM_12_3

[0164] (dispatch_point8+t≤dispatch_point 12 +16000)

[0165] OR(dispatch_point 12 +16000+t≤dispatch_point8)

[0166] Comparison between SM_8_2 and SM_12_3

[0167] (dispatch_point8+20000+t≤dispatch_point 12 +16000)

[0168] OR(dispatch_point 12 +16000+t≤dispatch_point8+20000)

[0169] Comparison between SM_8_2 and SM_12_4

[0170] (dispatch_point8+20000+t≤dispatch_point 12 +24000)

[0171] OR(dispatch_point 12 +24000+t≤dispatch_point8+20000)

[0172] Comparison between SM_8_2 and SM_12_5

[0173] (dispatch_point8+20000+t≤dispatch_point 12 +32000)

[0174] OR(dispatch_point 12 +32000+t≤dispatch_point8+20000)

[0175] Timing requirement constraints:

[0176] The timing constraints between VL_1 and VL_7 and VL_8 involve the Smessages of the sending links: SM_1, SM_32, and SM_37. The timing requirements are set as follows:

[0177] dispatch_point1+1000=dispatch_point 32

[0178] dispatch_point1+1500=dispatch_point 37

[0179] Initial auxiliary constraints:

[0180] For SM_1, its initial auxiliary constraints are:

[0181] dispatch_point1≥200

[0182] Incremental constraints:

[0183] Incremental constraints are used to maintain the result state of the previous plan during planning. Assume that the time when SM_1 is sent from ES_7 in the first plan is 1ms. Then, the following constraints should be set in the second and third plans:

[0184] dispatch_point1=1000

[0185] In summary:

[0186] (1) By dividing messages, designing constraints, and planning in stages, the message planning time can be shortened and the actual scale of network message services that can be planned can be increased. Figure 6 As shown ( Figure 6 The horizontal axis represents the message volume, and the vertical axis represents the solution time). When the scale of TT messages increases, the step-by-step planning method of the present invention can shorten the solution time;

[0187] (2) Message division and selection are automatically generated by link load calculation, without the need for manual intervention;

[0188] (3) Compared with the traditional SMT-based TTE message planning method that cannot directly obtain the solver status, since the messages are divided into three groups and planned in stages, the message planning progress can be observed by observing the result files generated by each planning and the software output.

Claims

1. A message increment planning method, characterized in that: The method steps are as follows: S1: Decompose the message Message into Smessage according to all the links that the corresponding virtual link passes through, and obtain the corresponding window size and minimum frame interval of Smessage; S2: Apply Smessage obtained in step S1 to divide all TT messages in the network according to the preset division rules, and determine whether all messages are divided into the specified set. If so, execute step S3; S3: Build the SMT solver, set the initial scope of the SMT solver, and perform the initial auxiliary constraints of the SMT solver; S4: Apply the set data obtained in step S2 and the initial auxiliary constraint SMT solver obtained in step S3, set corresponding constraints according to the corresponding sets, and generate a set-constraint SMT solution model; S5: Apply the set-constraint SMT solution model obtained in step S4, input the data of the corresponding set into the set-constraint SMT solution model to calculate and obtain the information planning result.

2. A message increment planning method according to claim 1, characterized in that: In step S1: S11: Obtain the message transmission path according to the virtual link VL of the message; S12: Apply the message transmission path obtained in step S11, and generate a corresponding Smessage for each physical link passed through the message transmission path; S13: Apply the Smessage obtained in step S12 and calculate the corresponding window size and minimum frame interval according to the Smessage transmission rate and message attributes.

3. A message increment planning method according to claim 1, characterized in that: In step S2: S21: Select messages with mandatory constraints on the sending timing from all TT messages, and decompose the selected messages to generate all Smessage sets as the first set vectorGNC of message planning; S22: Apply the process of filtering the remaining TT messages in step S21, analyze the remaining TT message loads, select the device port with the largest message load, and select all Smessage sets generated by decomposing the messages corresponding to all virtual links (VLs) passing through this link as the second set (vectorCongestion) of message planning; S23: Apply all the Smessage sets generated by decomposing the remaining messages after screening in step S21 and step S22 as the third set vectorOther of message planning.

4. A message increment planning method according to claim 3, characterized in that: In step S22: TTE network uses full-duplex network. For a message sending and receiving port,<devid,portid> To uniquely identify each Smessage, each Smessage also corresponds to a unique message sending and receiving port. Statistics are performed on the Smessages corresponding to the remaining TT messages filtered in step S21. The formula for calculating the load of each link is as follows: Where L represents the load of the link, window i Indicates the window time of the i-th Smessage on the current link, TTGap i Indicates the frame interval of the Smessage, clusterCycle is the cluster period of the TTE network, period i The message cycle corresponding to the Smessage.

5. A message increment planning method according to claim 1, characterized in that: In step S4, the constraints include: single cycle constraint, offset constraint, conflict-free constraint, message window constraint, timing requirement constraint and increment constraint; Define S as the set of Smessage variables for planning, sm is one of the Smessage variables, dispatch_point is the sending time point, window is the time window size, TTGap is the message interval, clusterCycle is the cluster period of the TTE network, period i is the message cycle corresponding to the i-th Smessage, and dispatch_point refers to its first sending, that is, the sending point within its message cycle; The single-cycle constraint is: The offset constraints are: The period of TT message is an integer multiple of the network integration period inteCycle: period=n×inteCycle,n∈Z,n>0 For each message there should be: Where period represents the message cycle, n is an integer indicating that the message cycle is a multiple of the integration cycle, inteCycle represents the integration cycle of network messages, k represents the kth integration cycle, shift_time represents the time period at the beginning of each integration cycle that cannot be used for message scheduling, and dispatch_point represents the time when the message is sent; The conflict-free constraints are: (dispatch_point i +window i +TTGap i +k×period i ≤dispatch_point j +l×period j ) OR(dispatch_point j +window j +TTGap j +l×period j ≤dispatch_point i +k×period i ) Where k represents the kth sending of message i, and l represents the lth sending of message j; The message window constraint is: set according to the message forwarding mechanism of the actual network device; Timing requirement constraints are: set according to actual message service requirements; The increment constraints are: Among them, pre_result i Indicates sm i Pre-planning results.

6. A message increment planning method according to claim 5, characterized in that: In step S4, the set-constraint SMT solution model includes: using single-cycle constraints, offset constraints, conflict-free constraints, message window constraints and timing requirement constraints for the data in the first set vectorGNC; using single-cycle constraints, offset constraints, conflict-free constraints, message window constraints and incremental constraints for the data in the second set vectorCongestion; using single-cycle constraints, offset constraints, conflict-free constraints, message window constraints and incremental constraints for the data in the third set vectorOther.

7. A message increment planning method according to claim 5, characterized in that: In step S3: The initial auxiliary constraints are:

Images