A precise prediction and multi-scenario coordination control system for container truck arrival time sequence

Abstract

The application discloses a precise prediction and multi-scene cooperative control system for container truck arrival time sequence, and relates to the technical field of port logistics scheduling. The system comprises an event propagation module, a time delay processing module, an inter-chain constraint module, an archive self-correction module and a deduction control module. The event propagation module constructs a causal propagation network and triggers associated event concomitant time delay. The time delay processing module performs non-accumulative logical operation on multiple time delays with causal correlation or time overlap based on a time delay fusion rule set. The inter-chain constraint module maintains a set of conditional constraint relationships between event chains. The archive self-correction module maintains an individual logical archive composed of logical state bits for each transportation unit, dynamically adjusts the logical state bits and ignores misjudgment event chains. The deduction control module obtains the individual logical archive, performs conditional verification with the logical state bits in deduction, executes inter-chain constraint rules and generates a time sequence prediction result.

Description

Technical Field

[0001] This invention relates to the field of port logistics scheduling technology, specifically a precise prediction and multi-scenario collaborative control system for the arrival time of container trucks at the port. Background Technology

[0002] Predicting the arrival time of container trucks at container terminals is a key input for terminal resource scheduling and operational planning. During the entire chain of truck travel from off-site roads and gates to the yard, there is a cascading propagation of various disturbance events. For example, queuing off-site can lead to gate congestion, which in turn can further affect the efficiency of yard operations, forming a complex causal chain of influence between events.

[0003] Existing container truck arrival forecasting solutions are mostly based on statistical modeling of historical data or queuing theory analysis in a single scenario, lacking a systematic design for handling causal propagation networks between multiple events and the nonlinear overlapping of time delays. Existing solutions typically employ a unified forecasting model oriented towards the group, failing to consider the differentiated behavioral characteristics of individual container trucks developed over long-term operation and their heterogeneous responses to disturbances, making accurate individualized time-series forecasting difficult. When multiple causal event chains cannot function simultaneously due to physical exclusivity or logical conflicts, existing solutions lack cross-chain dynamic constraints and individualized adjudication mechanisms. Summary of the Invention

[0004] The purpose of this invention is to provide an accurate prediction and multi-scenario collaborative control system for the arrival time of container trucks, so as to solve the problems raised in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a precise prediction and multi-scenario collaborative control system for the arrival time of container trucks, the system comprising: an event propagation module, a delay processing module, an inter-chain constraint module, an archive self-correction module, and a deduction control module; The event propagation module is used to construct a causal propagation network between events. When a basic event is identified, related events are triggered according to the propagation rules and accompanied by a delay. Related events generate further delays or reverse the delay of activated events based on their own rules. The delay processing module is used to determine the comprehensive delay value based on the delay fusion rule set, using non-cumulative logical operations on multiple delays that are generated during the causal propagation of events and have causal correlation or time overlap. The inter-chain constraint module is used to maintain the set of conditional constraint relationships between event chains and define the rules for mutual suppression or mutual weakening of different event chains under specific conditions. The file self-correction module is used to maintain an individual logical file consisting of multiple logical status bits for each transportation unit. After each task, it compares the actual impact of each event chain reflected by the actual arrival time with the event chain activated in this task and the constraint execution record, and adjusts each logical status bit accordingly and ignores misjudged event chains. The deduction control module is used to obtain the individual logical file of the current transportation unit, perform condition verification based on the logical state bits in the individual logical file at each step of the event propagation deduction, execute the inter-chain constraint rules corresponding to the current transportation unit, and determine activation or suppression based on the logical state bits of the current transportation unit when multiple event chains conflict due to constraints, and generate timing prediction results.

[0006] According to the above scheme, the event propagation module includes an event identification unit, a causal network construction unit, an event propagation unit, and a reverse inhibition unit; The event recognition unit is used to monitor and identify various basic events occurring in off-site, gate, and storage yard scenarios; The causal network construction unit is used to construct and maintain a causal propagation network between events, wherein the causal propagation network stores the propagation path, triggering conditions and inhibition relationships between each event. The event propagation unit is used to trigger related events according to the propagation rules in the causal network construction unit when the event identification unit identifies a basic event, and to generate the time delay associated with the related event at the same time as triggering the related event; The reverse inhibition unit is used to reverse inhibit the duration of the delay of the activated primary event based on the inhibition relationship stored in the causal network construction unit after the associated event is triggered.

[0007] According to the above scheme, the causal network construction unit includes: Off-site events, gate events, and yard events are defined as independent nodes in the causal propagation network, and a unique event identifier is assigned to each node. A directed propagation edge is constructed between nodes that have a causal relationship. The directed propagation edge includes a source node identifier, a target node identifier, a propagation direction, and a propagation triggering condition. Construct reverse suppression edges between nodes that have a suppression relationship. The reverse suppression edges include the identifier of the suppressed node, the identifier of the suppression source node, the suppression type, and the suppression effective condition. All nodes, directed propagation edges, and reverse suppression edges are stored in a structured manner to construct the causal propagation network.

[0008] According to the above scheme, the latency processing module includes an association judgment unit, a rule storage unit, and a fusion operation unit; The association judgment unit is used to receive multiple delays generated during the causal propagation of an event and to determine whether there is a causal relationship and whether there is time overlap between the multiple delays. The rule storage unit is used to store a latency fusion rule set, which defines fusion operation methods corresponding to different association types and overlapping states. The fusion operation unit is used to call the corresponding fusion operation method in the rule storage unit according to the judgment result of the association judgment unit, perform non-cumulative logical operations on each delay and output a comprehensive delay value.

[0009] According to the above scheme, the fusion computing unit includes: When there is a causal relationship and time overlap among the multiple delays to be merged, a dominant delay is selected from the multiple delays and a collaborative correction is applied. The result is used as the comprehensive delay value. When there is a causal relationship between multiple delays to be merged but their times do not overlap, mutual exclusion processing is performed on the multiple delays, and only a single delay value is retained as the comprehensive delay value; When multiple delays to be merged have no causal relationship but overlap in time, each delay is subjected to amplitude limiting processing and then merged, and the result is used as the comprehensive delay value. When multiple correlations and overlaps exist simultaneously in the same set of time delays, the corresponding fusion operation methods are executed in sequence according to the order of causal correlation taking precedence over non-causal correlation or time overlap taking precedence over time non-overlap.

[0010] According to the above scheme, the inter-chain constraint module includes a constraint storage unit, a trigger monitoring unit, and a constraint execution unit; The constraint storage unit is used to store the set of conditional constraint relationships between event chains. Each constraint rule is defined with a first event chain identifier, a second event chain identifier, a constraint type, and a triggering condition. The trigger monitoring unit is used to monitor the activation status of each event chain in real time. When an event chain is detected to be activated and the triggering condition in the corresponding constraint rule is met, a constraint execution instruction is generated. The constraint execution unit is used to impose restrictions on the activation state or influence of another event chain according to the constraint execution instruction and the corresponding constraint type.

[0011] According to the above scheme, the constraint types defined in the constraint storage unit include forced disabling and contribution weakening; forced disabling is used to prevent the target event chain from being activated in the current process; contribution weakening is used to reduce the participation weight of the delay generated after the target event chain is activated in the overall delay.

[0012] According to the above scheme, the archive self-correction module includes an archive storage unit, a record storage unit, and a comparison and correction unit; The file storage unit is used to maintain an individual logical file for each transportation unit. The individual logical file consists of multiple logical status bits, including sensitivity switch bits corresponding to each event chain and interlock tendency switch bits corresponding to the constraint rules between each chain. The recording storage unit is used to receive and record the sequence of event chains activated in the current process and all inter-chain constraint rules that have been executed; The comparison and correction unit is used to logically compare the actual impact of each event chain reflected by the actual arrival time with the records in the record storage unit after the current journey ends, and update each logical status bit in the file storage unit according to the comparison result.

[0013] According to the above scheme, the comparison and correction unit includes: For each event chain activated in the current itinerary, determine whether the event chain has actually occurred based on the actual arrival time. If it is determined that the event chain actually occurred but the predicted delay exceeds the actual impact range, then the sensitivity switch corresponding to the event chain is switched to the low value zone. If it is determined that the event chain did not actually occur or did not have a significant impact, the event chain is marked as falsely triggered, and the triggering of the event chain is conditionally ignored in the next trip. The conditional ignoring is effective in a preset number of consecutive trips, and the ignoring mark is automatically removed after the preset number of trips is exceeded, or the ignoring mark is removed in advance when the event chain is detected to actually occur again in any subsequent trip. For each inter-chain constraint rule executed in the current process, if it is determined afterward that the event chain suppressed by the constraint rule should have been triggered normally, then the interlock tendency switch corresponding to the constraint rule is switched to the tendency not to suppress state. If it is determined that the delay is repeatedly calculated due to the failure to execute a certain inter-chain constraint rule, then the interlock tendency switch corresponding to that constraint rule is switched to the forced enable state.

[0014] According to the above scheme, the simulation control module includes: Obtain the individual logical file corresponding to the current transportation unit, and read all logical status bits in the individual logical file; During the event propagation simulation, for each event chain to be activated, check whether the event chain is marked as ignored in the file. If it is marked as ignored, skip the activation. For event chains that are not ignored, the state of the sensitivity switch corresponding to the event chain is used as one of the conditions for the verification. Activation is only allowed after the verification is passed. When multiple event chains simultaneously meet the activation conditions but have a mutual exclusion relationship defined by inter-chain constraint rules, the interlock tendency switch bit corresponding to the relevant inter-chain constraint rules in the file is read, and the state of the switch bit determines which event chain to activate or whether to suppress multiple event chains that meet the conditions simultaneously; when the interlock tendency switch bit is in a neutral state, activation or suppression is determined according to the default priority rules, the default priority including prioritizing the event chain that first meets the activation conditions and suppressing the other event chains; By combining the latency and interlocking results of all activated event chains, the arrival time prediction result of the current transportation unit is generated.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention improves the consistency between prediction and real-world scenarios by constructing a causal propagation network between events, which triggers related events and generates time delays by propagating disturbances of various basic events in a chain-like manner. 2. This invention avoids the problem of repeated calculation or improper superposition of delays caused by algebraic accumulation by performing non-cumulative logical operations on multiple delays generated during the causal propagation of events, making the comprehensive delay value closer to reality; 3. This invention maintains an individual logic file for each transportation unit, consisting of sensitivity switch bits and interlocking tendency switch bits, enabling the system to evolve differentiated prediction logic for different trucks, thus achieving accurate arrival time prediction for each truck. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the structure of a precise prediction and multi-scenario collaborative control system for the arrival time of container trucks, as described in this invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Example: Figure 1 As shown, the present invention provides a technical solution, a precise prediction and multi-scenario collaborative control system for the arrival time of container trucks, the system comprising: an event propagation module, a delay processing module, an inter-chain constraint module, an archive self-correction module, and a deduction control module; The event propagation module is used to build a causal propagation network between events. When a basic event is identified, related events are triggered according to the propagation rules and accompanied by delays. Related events generate further delays based on their own rules or reverse the delay of activated events. The delay processing module is used to determine the comprehensive delay value based on the delay fusion rule set, using non-cumulative logical operations for multiple delays that are generated during the causal propagation of events and have causal correlation or time overlap. The inter-chain constraint module is used to maintain the set of conditional constraint relationships between event chains and define the rules for mutual suppression or mutual weakening of different event chains under specific conditions. The file self-correction module is used to maintain an individual logical file consisting of multiple logical status bits for each transportation unit. After each task, it compares the actual impact of each event chain reflected by the actual arrival time with the event chain activated and constraint execution records in this task, and adjusts each logical status bit accordingly and marks the erroneous event chain as ignored. The deduction control module is used to obtain the individual logical file of the current transportation unit, perform condition verification based on the logical state bits in the individual logical file at each step of the event propagation deduction, execute the inter-chain constraint rules corresponding to the current transportation unit, and determine activation or suppression based on the logical state bits of the current transportation unit when multiple event chains conflict due to constraints, and generate timing prediction results.

[0019] Specifically, the event propagation module includes an event identification unit, a causal network construction unit, an event propagation unit, and a reverse suppression unit. The event identification unit is used to monitor and identify various basic events occurring in off-site, gate, and yard scenarios. The causal network construction unit is used to construct and maintain a causal propagation network between events, storing the propagation paths, triggering conditions, and suppression relationships between each event. The event propagation unit is used to trigger related events according to the propagation rules in the causal network construction unit when the event identification unit identifies a basic event, and to generate the delay associated with the related event at the same time. The reverse suppression unit is used to reverse suppress the duration of the delay of the activated primary event according to the suppression relationships stored in the causal network construction unit after the related event is triggered.

[0020] Specifically, this embodiment uses a single transportation trip as an example to illustrate the workflow of the event propagation module; the preset suppression ratio coefficient is 0.3, meaning the delay after reverse suppression is no less than 70% of the original delay; the preset minimum retention ratio is 0.5; in this trip, the event identification unit successively identifies off-site light queuing events, gate traffic short-term aggregation events, and yard shift change buffer overflow events; the causal network construction unit defines the above three types of events as independent nodes in the causal propagation network, and assigns a unique event identifier to each node, specifically off-site queuing node N1, gate aggregation node N2, and gate aggregation node N3. N2 is the overflow node in the storage yard, and N3 is the overflow node. Directed propagation edges between nodes include: a directed propagation edge from N1 to N2, triggered when the queuing time exceeds 3 minutes; and a directed propagation edge from N2 to N3, triggered when the aggregation duration exceeds 5 minutes. Reverse suppression edges include: a reverse suppression edge from N2 to N1, with suppression type being delay reduction. Suppression takes effect immediately after N2 is activated, and the suppression magnitude is limited by a preset suppression ratio coefficient of 0.3, meaning the suppressed delay is not less than 70% of the original delay. When the event propagation unit receives information about N1, it generates [something] according to the propagation rules. The associated delay value of N1 is 25 minutes, and the propagation triggering condition is checked; the queuing time is determined to be 4 minutes, which meets the propagation triggering condition of N1 pointing to N2. The event propagation unit then activates N2 and generates an associated delay value of 18 minutes for N2; after N2 is activated, the reverse suppression unit performs a reverse suppression operation on the delay duration of N1 according to the reverse suppression edge. In this embodiment, the formula for reverse suppression is: T_suppressed=max(T_original×(1-β), T_original×γ); where T_suppressed d is the delay value after reverse suppression, T_original is the original delay value of the primary event, β is the preset suppression ratio coefficient, and γ is the preset minimum retention ratio; in this embodiment, β is 0.3 and γ is 0.5; T_suppressed = max(25 × (1-0.3), 25 × 0.5) = 17.5 minutes is obtained; after N2 is activated, the gate aggregation duration is 6 minutes, which satisfies the propagation trigger condition of N2 pointing to N3. The event propagation unit further activates N3 and generates the associated delay value of N3, which is 12 minutes. This is only an example and is not a limitation.

[0021] Specifically, the building blocks of causal networks include: Off-site events, gate events, and yard events are defined as independent nodes in the causal propagation network, and each node is assigned a unique event identifier. Directed propagation edges are constructed between nodes with causal relationships, and each directed propagation edge includes the source node identifier, the target node identifier, the propagation direction, and the propagation triggering condition. Reverse inhibition edges are constructed between nodes with inhibition relationships, and each reverse inhibition edge includes the inhibited node identifier, the inhibition source node identifier, the inhibition type, and the inhibition effective condition. All nodes, directed propagation edges, and reverse inhibition edges are stored in a structured manner to construct the causal propagation network.

[0022] Specifically, in this embodiment, the causal network construction unit defines the off-site light queuing event, the gate traffic congestion event, and the yard shift change buffer overflow event as independent nodes in the causal propagation network, and assigns a unique event identifier to each node: off-site queuing node N1, gate congestion node N2, and yard overflow node N3. The directed propagation edge between nodes includes a directed propagation edge from N1 to N2, with the source node identified as N1, the target node identified as N2, the propagation direction as from N1 to N2, and the propagation trigger condition as a queuing time exceeding 3 minutes. The directed propagation edge from N2 to N3 has N2 as the source node and N3 as the target node. The propagation direction is from N2 to N3, and the propagation trigger condition is that the aggregation duration exceeds 5 minutes. The reverse suppression edge includes the reverse suppression edge from N2 to N1, with N1 as the suppressed node and N2 as the suppression source node. The suppression type is time delay reduction, and the suppression takes effect immediately after N2 is activated. All nodes, directed propagation edges, and reverse suppression edges are stored in a structured form to form a causal propagation network. This is only an example and is not a limitation.

[0023] Specifically, the latency processing module includes an association judgment unit, a rule storage unit, and a fusion operation unit. The association judgment unit is used to receive multiple latencys generated during the causal propagation of events and determine whether there is a causal relationship or time overlap between the multiple latencys. The rule storage unit is used to store a latency fusion rule set, which defines fusion operation methods corresponding to different association types and overlap states. The fusion operation unit is used to call the corresponding fusion operation method in the rule storage unit based on the judgment result of the association judgment unit, perform non-cumulative logical operations on each latency, and output a comprehensive latency value.

[0024] Specifically, after the causal propagation process of the event ends, three delay values ​​are generated: N1 delay of 17.5 minutes, N2 delay of 18 minutes, and N3 delay of 12 minutes after reverse suppression. These three delays are sent to the delay processing module. The association judgment unit receives these three delays and judges the association type and overlap state between each pair. In this embodiment, the judgment rule adopts the time sequence relationship judgment algorithm. The time intervals of the two delays are [T1_start, T1_end] and [T2_start, T2_end], respectively. If T1_end ≥ T2_start and T2_end ≥ T1_start, then it is judged as time overlap; otherwise, it is judged as time non-overlap. The causal association judgment criteria include: if the nodes corresponding to the two delays are connected by directed propagation edges in the causal propagation network, then it is judged as having a causal association; otherwise, it is judged as not having a causal association. The following conditions are met: 1. No causal relationship is determined. 2. Causal relationship is determined. 3. A causal relationship exists between delay N1 and delay N2, i.e., there is a directed propagation edge from N1 to N2, and the time intervals of the two types of events overlap, which is a case of causal relationship and time overlap. 4. A causal relationship exists between delay N2 and delay N3, i.e., there is a directed propagation edge from N2 to N3, but the occurrence time of the overflow event is later than the end time of the gate accumulation event, which is a case of causal relationship but no time overlap. 5. There is no direct directed propagation edge connecting delay N1 and delay N3, but their time intervals partially overlap, which is a case of no causal relationship but time overlap. 6. Three different combinations of relationships and overlaps exist simultaneously in this group of delays. The fusion operation unit processes them according to a prescribed priority order. In this embodiment, the priority order includes: causal relationship takes precedence over non-causal relationship, and time overlap takes precedence over no time overlap.

[0025] Specifically, the N1 and N2 delays, which are causally related and overlap in time, are processed first. The fusion unit selects the larger N2 delay (18 minutes) as the dominant delay and applies a collaborative correction to it. In this embodiment, the collaborative correction uses a dominant delay weighted correction algorithm, with the formula: T_fused = T_dominant × (1 + α × (1 - T_non-dominant / T_dominant)); where T_fused represents the corrected combined delay value; and T_dominant represents the selected dominant delay. The value is 18 minutes; T_non-dominant represents another unselected delay value, which is 17.5 minutes; α is a preset co-correction coefficient, which is 0.15 in this embodiment; T_fused = 18 × (1 + 0.15 × (1 - 17.5 / 18)) ≈ 18.08 minutes; then process the causally related but non-overlapping delays N2 and N3; the N2 delay has been fused, and the fusion operation unit performs mutual exclusion processing, retaining only the fused delay value of N1 and N2 after synthesis, which is 18.08 minutes, and discarding the N2 delay value of 18 minutes alone. Next, process the N1 and N3 delays that have no causal relationship but overlap in time; the N1 delay has been fused, and the N3 delay is still 12 minutes; the fusion operation unit performs amplitude limiting processing on each delay, including comparing each delay value with a preset saturation threshold of 40 minutes. If it exceeds the threshold, it is truncated to the threshold; otherwise, the original value is retained; here, neither delay value exceeds the 40-minute threshold, so no truncation is required; after amplitude limiting, the two delays are merged. In this embodiment, the merging adopts the amplitude limiting weighted merging algorithm, and the formula is: T_merged=Σmin(T_i,T_sat)×w_i; Where T_merged is the merged overall delay value; T_i is the original value of the i-th delay to be merged; T_sat is the preset saturation threshold, which is 40 minutes here; w_i is the merging weight corresponding to the i-th delay, and in this embodiment, the weight of each delay is 1.0; T_merged = min(18.08,40)×1.0 + min(12,40)×1.0 = 30.08 minutes is obtained; the overall delay value is 30.08 minutes, which is the delay prediction result of the causal propagation of this event. This is only an example and is not a limitation.

[0026] Specifically, the fusion computing unit includes: When multiple delays to be fused are causally related and time-overlapping, a dominant delay is selected from the multiple delays and a collaborative correction is applied. The result is used as the comprehensive delay value. When multiple delays to be fused are causally related but time-overlapping, mutual exclusion processing is performed on the multiple delays, and only a single delay value is retained as the comprehensive delay value. When multiple delays to be fused are not causally related but time-overlapping, each delay is subjected to amplitude limiting processing and then merged. The result is used as the comprehensive delay value. When multiple combinations of correlation and overlap exist in the same group of delays, the corresponding fusion operation is performed in sequence according to the order of causal correlation taking precedence over non-causal correlation or time overlap taking precedence over time non-overlap.

[0027] Specifically, the inter-chain constraint module includes a constraint storage unit, a trigger monitoring unit, and a constraint execution unit. The constraint storage unit stores the set of conditional constraint relationships between event chains. Each constraint rule defines a first event chain identifier, a second event chain identifier, a constraint type, and a trigger condition. The trigger monitoring unit monitors the activation status of each event chain in real time. When an event chain is detected to be activated and meets the trigger condition in the corresponding constraint rule, a constraint execution instruction is generated. The constraint execution unit imposes restrictions on the activation status or influence of another event chain based on the constraint execution instruction and the corresponding constraint type.

[0028] Specifically, the constraint storage unit of the inter-chain constraint module stores multiple conditional constraint rules. In this embodiment, the constraint rules include: Rule R1, which defines the first event chain as event chain N1 to N2 and the second event chain as event chain N1 to N3, with a constraint type of forced disabling and a trigger condition of when event chain N1 to N2 is activated; and Rule R2, which defines the first event chain as event chain N2 to N3, with a constraint type of contribution weakening and a weakening magnitude of 0.5. The trigger monitoring unit monitors the activation status of each event chain in real time; when it detects that event chain N1 to N2 is activated... When active, a first constraint execution instruction corresponding to rule R1 is generated; when the activation of the N2 to N3 event chain is detected, a second constraint execution instruction corresponding to rule R2 is generated; the constraint execution unit applies an activation restriction to the N1 to N3 event chain according to the first constraint execution instruction and the forced disable constraint type, so that it is not activated in the current process; according to the second constraint execution instruction and the contribution weakening constraint type, the participation weight of the delay generated after the N2 to N3 event chain is activated in the overall delay is reduced to 0.5 times the original weight. This is only an example and is not a limitation.

[0029] Specifically, the constraint types defined in the constraint storage unit include forced disable and contribution weakening; forced disable is used to prevent the target event chain from being activated in the current process; contribution weakening is used to reduce the participation weight of the latency generated after the target event chain is activated in the overall latency.

[0030] Specifically, the file self-correction module includes a file storage unit, a record storage unit, and a comparison and correction unit. The file storage unit is used to maintain an individual logical file for each transportation unit. The individual logical file consists of multiple logical status bits, including sensitivity switch bits corresponding to each event chain and interlocking tendency switch bits corresponding to the inter-chain constraint rules. The record storage unit is used to receive and record the event chain sequence activated in the current trip and all inter-chain constraint rules executed. The comparison and correction unit is used to logically compare the actual impact of each event chain reflected by the actual arrival time with the records in the record storage unit after the current trip ends, and update each logical status bit in the file storage unit according to the comparison results.

[0031] Specifically, after this trip, the storage unit has fully recorded: the activated event chain sequences are event chains N1 to N2 and N2 to N3; all executed inter-chain constraint rules are rules R1 and R2; the actual arrival time is compared with the predicted arrival time. In this trip, the predicted arrival time is the base time + 30.08 minutes, while the actual arrival time is the base time + 22 minutes; for each activated event chain, the comparison and correction unit determines whether it actually occurred based on the actual arrival time; in this embodiment, the judgment method adopts the latency contribution back-calculation algorithm, where the actual total latency is T_actual, the predicted total latency is T_predict, and the predicted latency contribution of a certain event chain is T_chain. Then, the estimated actual contribution of the event chain is T_chain × (T_actual / T_predict); when the estimated actual contribution of a certain event chain is less than 20% of its predicted latency contribution, it is determined that the event chain did not actually occur. Or it may not have a significant impact; the actual contribution estimate of the N1 to N2 event chain is 18.08×(22 / 30.08)≈13.22 minutes, which is higher than 18.08×20%=3.62 minutes, and is judged to have actually occurred. However, its predicted delay exceeds the actual impact range. The sensitivity switch corresponding to this event chain is switched to the low value zone, and the direction of the low value zone is adjusted towards the OFF direction. The delay prediction weight of the corresponding event chain is reduced and the trigger threshold is tightened. In this embodiment, the sensitivity switch... The adjustment adopts a step adjustment algorithm. When the switch position is in a multi-value state, moving to the low value zone means subtracting the preset step size from the current value. In this embodiment, the preset step size is 1 level, and the direction of the low value zone is to adjust one level towards the OFF direction. The actual contribution estimate of the N2 to N3 event chain is 12×(22 / 30.08)×0.5=4.39 minutes, and the predicted delay contribution of this event chain is 12×0.5=6 minutes. The actual contribution estimate of 4.39 minutes is higher than 6×20%=1.Two minutes are required for an event chain to be considered a true occurrence. If an event chain is determined not to have actually occurred during a given run, it is marked as a false trigger and conditionally ignored in the next run. The ignore flag is valid for three consecutive runs, automatically removed after three runs, or removed early if the event chain is detected to have actually occurred again in any subsequent run. For the inter-chain constraint rules executed in this run, rule R1, which forcibly disables event chains N1 to N3, is subsequently determined to be a suppressed event chain that was not triggered in this run and requires no adjustment. Rule R2, which contributes to weakening event chains N2 to N3, is subsequently determined to be a suppressed event chain. If the weakened event chain actually occurred and its actual contribution matches the weakened contribution, no adjustment is needed. If it is subsequently determined that the event chain suppressed by the constraint rule should have triggered normally, then the interlock tendency switch corresponding to that constraint rule is switched to the tendency not to suppress state. If delays are recalculated due to the failure to execute a constraint rule, then the interlock tendency switch corresponding to that constraint rule is switched to the forced enable state. The adjustment of the interlock tendency switch also uses a step-by-step adjustment algorithm, adjusting one level at a time. Possible values ​​include forced enable, tendency enable, neutral, tendency not to suppress, and forced disable. This is only an example and is not a limitation.

[0032] Specifically, the comparison and correction unit includes: For each event chain activated in the current trip, determine whether the event chain actually occurred based on the actual arrival time. If it is determined to have actually occurred but the predicted delay associated with the event chain exceeds the actual impact range, then the sensitivity switch corresponding to the event chain is moved to a low value zone. If it is determined not to have actually occurred or not to have a significant impact, then the event chain is marked as falsely triggered, and its triggering is conditionally ignored in the next trip. Conditional ignoring is effective for a preset number of consecutive trips, and the ignoring mark is automatically removed after the preset number of trips, or the ignoring mark is removed in advance when the event chain is detected to have actually occurred again in any subsequent trip. For each inter-chain constraint rule executed in the current trip, if it is subsequently determined that the event chain suppressed by the constraint rule should have been triggered normally, then the interlock tendency switch corresponding to the constraint rule is moved to the non-suppression state. If it is determined that the delay is recalculated due to the failure to execute an inter-chain constraint rule, then the interlock tendency switch corresponding to the constraint rule is moved to the forced enable state.

[0033] Specifically, the simulation control module includes: Obtain the individual logical file corresponding to the current transportation unit and read all logical state bits in the individual logical file. During the event propagation and deduction process, for each event chain to be activated, check whether the event chain is marked as ignored in the file. If it is marked as ignored, skip the activation. For event chains that are not ignored, use the state of the sensitivity switch bit corresponding to the event chain as one of the conditions for verification. Activation is only allowed after the verification passes. When multiple event chains simultaneously meet the activation conditions but there is a mutual exclusion relationship defined by the inter-chain constraint rules, read the interlock tendency switch bit corresponding to the relevant inter-chain constraint rules in the file. Determine which event chain to activate or whether to suppress multiple event chains that meet the conditions simultaneously based on the state of the switch bit. When the interlock tendency switch bit is in a neutral state, determine whether to activate or suppress according to the default priority rules. The default priority includes prioritizing the event chain that first meets the activation conditions and suppressing the other event chains. Combine the delays and interlock processing results of all activated event chains to generate the arrival time prediction result of the current transportation unit.

[0034] Specifically, in the simulation and prediction of the next trip, the simulation control module obtains the individual logic file corresponding to the current transportation unit and reads all the logic status bits in the file; during the event propagation simulation, for each event chain to be activated, it checks whether it has been marked as ignored; the N1 to N3 event chain in the previous trip was not marked as ignored, so the check continues; for the event chain that has not been ignored, the state of the sensitivity switch bit corresponding to the N1 to N2 event chain is used as one of the activation verification conditions; since the switch bit was switched to the low value area in the previous trip, the activation condition verification of the event chain in this simulation will effectively increase the queuing time threshold in the activation condition, that is, the queuing time judgment condition required to trigger the event chain is stricter; When multiple event chains simultaneously meet the activation conditions but have mutual exclusion relationships, the interlocking tendency switch corresponding to the relevant inter-chain constraint rules is read. If the interlocking tendency switch of a certain transport unit is in a neutral state, the decision is made according to the default priority rule. The default priority is to retain the event chain that first meets the activation conditions and suppress the other event chains. In this embodiment, the rule execution adopts a priority decision algorithm. The event chain set is C, where the timestamp of each event chain meeting the activation conditions is t_i. The mutual exclusion constraint relationship requires that only one event chain be retained from the mutual exclusion pair. For each event chain with mutual exclusion relationship, they are arranged in ascending order according to the activation timestamp t_i. The event chain with the smallest t_i is retained, and the other mutual exclusion event chains are suppressed. After completing all event propagation, delay fusion, inter-chain constraint and inference control steps, the arrival time prediction result of the current transport unit is generated by combining the delay and interlocking processing results of all activated event chains.

[0035] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A precise prediction and multi-scenario collaborative control system for the arrival time of container trucks, characterized in that: The system includes: an event propagation module, a delay processing module, an inter-chain constraint module, an archive self-correction module, and a deduction control module; The event propagation module is used to construct a causal propagation network between events. When a basic event is identified, related events are triggered according to the propagation rules and accompanied by a delay. Related events generate further delays or reverse the delay of activated events based on their own rules. The delay processing module is used to determine the comprehensive delay value based on the delay fusion rule set, using non-cumulative logical operations on multiple delays that are generated during the causal propagation of events and have causal correlation or time overlap. The latency processing module includes an association judgment unit, a rule storage unit, and a fusion operation unit; the fusion operation unit is used to call the corresponding fusion operation method in the rule storage unit according to the judgment result of the association judgment unit, perform non-cumulative logical operations on each latency, and output a comprehensive latency value. The fusion computing unit includes: When there is a causal relationship and time overlap among the multiple delays to be merged, a dominant delay is selected from the multiple delays and a collaborative correction is applied. The result is used as the comprehensive delay value. When there is a causal relationship between multiple delays to be merged but their times do not overlap, mutual exclusion processing is performed on the multiple delays, and only a single delay value is retained as the comprehensive delay value; When multiple delays to be merged have no causal relationship but overlap in time, each delay is subjected to amplitude limiting processing and then merged, and the result is used as the comprehensive delay value. When multiple combinations of correlation and overlap exist simultaneously in the same set of time delays, the corresponding fusion operation methods are executed in sequence according to the order of causal correlation taking precedence over non-causal correlation or time overlap taking precedence over time non-overlap. The inter-chain constraint module is used to maintain the set of conditional constraint relationships between event chains and define the rules for mutual suppression or mutual weakening of different event chains under specific conditions. The inter-chain constraint module further includes a constraint storage unit; the constraint storage unit is used to store a set of conditional constraint relationships between event chains, and each constraint rule defines a first event chain identifier, a second event chain identifier, a constraint type, and a triggering condition; the constraint types defined in the constraint rules of the constraint storage unit include forced disabling and contribution weakening; forced disabling is used to prevent the target event chain from being activated in the current process; contribution weakening is used to reduce the participation weight of the delay generated after the target event chain is activated in the overall delay; The file self-correction module is used to maintain an individual logical file consisting of multiple logical status bits for each transportation unit. After each task, it compares the actual impact of each event chain reflected by the actual arrival time with the event chain activated in this task and the constraint execution record, and adjusts each logical status bit accordingly and ignores misjudged event chains. The deduction control module is used to obtain the individual logical file of the current transportation unit, perform condition verification based on the logical state bits in the individual logical file at each step of the event propagation deduction, execute the inter-chain constraint rules corresponding to the current transportation unit, and determine activation or suppression based on the logical state bits of the current transportation unit when multiple event chains conflict due to constraints, and generate timing prediction results.

2. The accurate prediction and multi-scenario collaborative control system for container truck arrival time according to claim 1, characterized in that: The event propagation module includes an event identification unit, a causal network construction unit, an event propagation unit, and a reverse inhibition unit; The event recognition unit is used to monitor and identify various basic events occurring in off-site, gate, and storage yard scenarios; The causal network construction unit is used to construct and maintain a causal propagation network between events, wherein the causal propagation network stores the propagation path, triggering conditions and inhibition relationships between each event. The event propagation unit is used to trigger related events according to the propagation rules in the causal network construction unit when the event identification unit identifies a basic event, and to generate the time delay associated with the related event at the same time as triggering the related event; The reverse inhibition unit is used to reverse inhibit the duration of the delay of the activated primary event based on the inhibition relationship stored in the causal network construction unit after the associated event is triggered.

3. The accurate prediction and multi-scenario collaborative control system for container truck arrival time according to claim 2, characterized in that: The causal network construction unit includes: Off-site events, gate events, and yard events are defined as independent nodes in the causal propagation network, and a unique event identifier is assigned to each node. A directed propagation edge is constructed between nodes that have a causal relationship. The directed propagation edge includes a source node identifier, a target node identifier, a propagation direction, and a propagation triggering condition. Construct reverse suppression edges between nodes that have a suppression relationship. The reverse suppression edges include the identifier of the suppressed node, the identifier of the suppression source node, the suppression type, and the suppression effective condition. All nodes, directed propagation edges, and reverse suppression edges are stored in a structured manner to construct the causal propagation network.

4. The accurate prediction and multi-scenario collaborative control system for container truck arrival time according to claim 1, characterized in that: The association judgment unit is used to receive multiple delays generated during the causal propagation of an event and to determine whether there is a causal relationship and whether there is time overlap between the multiple delays. The rule storage unit is used to store the latency fusion rule set, which defines fusion operation methods corresponding to different association types and overlapping states.

5. The accurate prediction and multi-scenario collaborative control system for truck arrival timing according to claim 1, characterized in that: The inter-chain constraint module also includes a trigger monitoring unit and a constraint execution unit; The trigger monitoring unit is used to monitor the activation status of each event chain in real time. When an event chain is detected to be activated and the triggering condition in the corresponding constraint rule is met, a constraint execution instruction is generated. The constraint execution unit is used to impose restrictions on the activation state or influence of another event chain according to the constraint execution instruction and the corresponding constraint type.

6. The accurate prediction and multi-scenario collaborative control system for container truck arrival time according to claim 1, characterized in that: The self-correction module for archives includes an archive storage unit, a record storage unit, and a comparison and correction unit. The file storage unit is used to maintain an individual logical file for each transportation unit. The individual logical file consists of multiple logical status bits, including sensitivity switch bits corresponding to each event chain and interlock tendency switch bits corresponding to the constraint rules between each chain. The recording storage unit is used to receive and record the sequence of event chains activated in the current process and all inter-chain constraint rules that have been executed; The comparison and correction unit is used to logically compare the actual impact of each event chain reflected by the actual arrival time with the records in the record storage unit after the current journey ends, and update each logical status bit in the file storage unit according to the comparison result.

7. The accurate prediction and multi-scenario collaborative control system for truck arrival timing according to claim 6, characterized in that: The comparison and correction unit includes: For each event chain activated in the current itinerary, determine whether the event chain has actually occurred based on the actual arrival time. If it is determined that the event chain actually occurred but the predicted delay exceeds the actual impact range, then the sensitivity switch corresponding to the event chain is switched to the low value zone. If it is determined that the event chain did not actually occur or did not have a significant impact, the event chain is marked as falsely triggered, and the triggering of the event chain is conditionally ignored in the next trip. The conditional ignoring is effective in a preset number of consecutive trips, and the ignoring mark is automatically removed after the preset number of trips is exceeded, or the ignoring mark is removed in advance when the event chain is detected to actually occur again in any subsequent trip. For each inter-chain constraint rule executed in the current process, if it is determined afterward that the event chain suppressed by the constraint rule should have been triggered normally, then the interlock tendency switch corresponding to the constraint rule is switched to the tendency not to suppress state. If it is determined that the delay is repeatedly calculated due to the failure to execute a certain inter-chain constraint rule, then the interlock tendency switch corresponding to that constraint rule is switched to the forced enable state.

8. The accurate prediction and multi-scenario collaborative control system for truck arrival timing according to claim 1, characterized in that: The deduction control module includes: Obtain the individual logical file corresponding to the current transportation unit, and read all logical status bits in the individual logical file; During the event propagation simulation, for each event chain to be activated, check whether the event chain is marked as ignored in the file. If it is marked as ignored, skip the activation. For event chains that are not ignored, the state of the sensitivity switch corresponding to the event chain is used as one of the conditions for the verification. Activation is only allowed after the verification is passed. When multiple event chains simultaneously meet the activation conditions but have a mutual exclusion relationship defined by inter-chain constraint rules, the interlock tendency switch bit corresponding to the relevant inter-chain constraint rules in the file is read, and the state of the switch bit determines which event chain to activate or whether to suppress multiple event chains that meet the conditions simultaneously; when the interlock tendency switch bit is in a neutral state, activation or suppression is determined according to the default priority rules, the default priority including prioritizing the event chain that first meets the activation conditions and suppressing the other event chains; By combining the latency and interlocking results of all activated event chains, the arrival time prediction result of the current transportation unit is generated.

Images