Search

CN-121983261-A - Outpatient service queue AI decision bus method and system based on dynamic fusing and self-healing

CN121983261ACN 121983261 ACN121983261 ACN 121983261ACN-121983261-A

Abstract

The invention discloses an outpatient queue AI decision bus method and a system based on dynamic fusing and self-healing, wherein the method comprises the following steps: and S1, data acquisition and preprocessing, namely acquiring state, sorting, priority, load and check-in rule data from a patient table and an associated table. According to the invention, automatic degradation and service continuity under an abnormal scene are realized through the health monitoring and fuse state machine, self-adaptive coordination of multi-source rules is achieved by means of unified health evaluation and weight scheduling mechanism, factors such as fairness punishment, priority crowd and the like are quantified by means of an AI multi-factor scoring model to balance fairness and efficiency, and the strategy recovery compound rule controllability is ensured by shadow parallel evaluation and full-flow log audit, meanwhile, the existing system data table and SQL are multiplexed, the existing interface is compatible, and the hospital deployment cost is reduced on the basis of enhancing system robustness, improving decision science, guaranteeing service fairness and meeting supervision compliance.

Inventors

  • DOU JIANLIN
  • WANG YUSEN

Assignees

  • 厦门狄耐克物联智慧科技有限公司

Dates

Publication Date
20260505
Application Date
20260127

Claims (8)

  1. 1. The outpatient queue AI decision bus method based on dynamic fusing and self-healing is characterized by comprising the following steps: S1, data acquisition and preprocessing, namely acquiring state, sorting, priority, load and sign-in rule data from a patient table and an associated table; S2, health degree monitoring, namely constructing a health probe for a check-in switch, a display number strategy, load statistics and automatic diagnosis marking, and generating health degree H and fault indexes; s3, fusing judgment, namely maintaining a fuse state machine aiming at each rule/data source, and triggering a degradation strategy when abnormality occurs; S4, selecting a strategy set, namely selecting a main strategy set according to the current health state, wherein the main strategy set comprises a display number, a priority constraint, fairness punishment and load balancing weight; s5, AI multi-factor scoring, namely scoring and sorting candidate patients, outputting a queue sequence and a display number, and simultaneously reserving a shadow strategy for parallel scoring; S6, shadow evaluation, namely counting the difference of a shadow strategy and a main strategy on key indexes, and providing evidence for semi-open heuristic/restoration closure; S7, half-open trial and self-healing, namely switching the small-proportion flow to an alternative strategy when the shadow strategy counted in S6 shows stability, and observing the stability and then recovering to be closed; S8, audit record and visualization, namely recording factor constitution, weight, fusing state and switching reason of each decision, and supporting traceability and supervision; s9, online learning and weight self-adaption, namely automatically adjusting the weight and the threshold value of each factor based on the historical evaluation data; S10, outputting the existing queue, and accessing the existing interface and page in a compatible mode to ensure backward compatibility.
  2. 2. The dynamic fusing and self-healing based outpatient queue AI decision bus method as set forth in claim 1, wherein in S2, the health degree H is calculated using the formula H=1- α. FAILRATEI- β. TimeoutRatei- γ. MISSINGRATEI- δ. ConflictRatei; Wherein FAILRATEI is error rate, timeoutRatei is timeout rate, MISSINGRATEI is critical field missing rate, conflictRatei is rule conflict rate, and α, β, γ and δ are configurable weights.
  3. 3. The outpatient queue AI decision bus method based on dynamic blowing and self-healing according to claim 1, wherein in S3, the conditions of the blowing judgment are as follows: (1) If H < θopen or continuous K times trigger abnormality, the fuse state machine enters OPEN; (2) HALF-OPEN condition, namely if the shadow evaluation stability index Mshadow is more than or equal to theta shadow and H rises back to exceed delta H in the observation window, entering HALF-OPEN; (3) And closing conditions, namely restoring CLOSED when the probing is successful in the semi-open quantity proportion r and the abnormal times are less than or equal to epsilon.
  4. 4. The method of claim 1, wherein in S5, the following formula is used when performing multi-factor marking: S(p)=w1·Ftime(p)+w2·Fpriority(p)+w3·Ffairness(p)+w4·Fload(p)+w5·Fsign(p)-w6·Flock(p); Wherein Ftime is a time/sequence number factor, fpriority is a priority factor, FFAIRNESS is a fairness factor, flow is a load balancing factor, fsign is a check-in constraint factor, checked-in patients are marked when check-in is enabled, non-checked-in patients are punished or directly constraint filtered, flow is a locking factor, and is_lock_type=1 patient freezing position or punishment is performed.
  5. 5. The method for determining an AI decision bus of an outpatient queue based on dynamic blowing and self-healing as set forth in claim 1, wherein in S4, the display number policy is selected under the condition that when a display number rule is blown, a standby display number is used for degradation, and in a blown state, an unavailable main serial number is automatically backed.
  6. 6. The out-patient queue AI decision bus method based on dynamic fusing and self-healing according to claim 3, wherein in S6, a formula adopted by the shadow evaluation stability index is as follows: the shadow policy score ：Mshadow=η1·Norm(-AvgWaitTime)+η2·Norm(-OverRate)+η3·Norm(Throughput)+η4·Norm(Fairness)+η5·Norm(Balance), reaches the threshold θshadow before entering a half-open heuristic.
  7. 7. An outpatient queue AI decision bus system based on dynamic fusing and self-healing, which is used for realizing the method of any of claims 1-6, and is characterized by comprising a data acquisition and feature construction module, a rule and data health monitoring module, a fusing and self-healing management module, an AI decision engine module, a shadow evaluation module, a log audit and visualization module and a strategy management and configuration module; The data acquisition and feature construction module is used for butting out_ sick _info, qcs_visual_info and qcs_rule_template_info tables; The rule and data health monitoring module is used for establishing a health probe for check-in constraint, display number rule, load statistics and a consulting room start-stop state; The fusing and self-healing management module comprises a multi-fuse state machine, wherein the multi-fuse state machine enters OPEN in an abnormal state, and shadow evaluation successfully restores to be closed through a second half-OPEN heuristic; The AI decision engine module is used for multi-factor marking and constraint solving, and selecting different weights for different health states; The shadow evaluation module calculates scoring results of the standby strategy in parallel, and does not influence on-line and is used as a recovery closing basis; the log audit and visualization module is used for recording details of each decision, and assisting quality control and compliance; the strategy management and configuration module is used for visualizing configuration weights, thresholds, heuristic proportions and recovery criteria.
  8. 8. The dynamic-fuse and self-healing-based outpatient queue AI decision bus system of claim 7, wherein the logic of the rule and data health monitoring module is to establish a health H for check-in probes isEnableSignIn, display number selection serial number type, load data selectDoctorQueueCount, clinic start and stop getConsultationRoomQueueNumber; The fusing and self-healing management module is used for maintaining a state machine field and a time stamp of each rule_code and recovering the state machine field and the time stamp in linkage with shadow evaluation; The input of the AI decision engine module is a candidate patient list, and the process comprises multi-factor scoring/learning ordering+constraint+display number degradation logic, and the output is an ordered list and displayNumber; the shadow evaluation module calculates scores in parallel with the main strategy, does not change online sequencing, and records evaluation indexes; the log and audit module interfaces with the controller/service layer.

Description

Outpatient service queue AI decision bus method and system based on dynamic fusing and self-healing Technical Field The invention relates to the technical field of intelligent outpatient service, in particular to an outpatient service queue AI decision bus method and system based on dynamic fusing and self-healing. Background With popularization of hospital informatization systems, queuing number calling/triage/guide diagnosis capability in outpatient and medical technology scenes becomes a key of experience and efficiency, the conventional system generally adopts a rule-driven queue management model, and performs sequencing and scheduling by combining a check-in switch, display number selection, priority crowd, manual insertion and locking, diagnosis room stop/load state and the like, and the conventional system is taken as an example, and has typical structures and rules of multiple serial numbers and display, strategy coexistence, check-in constraint, multiple state mechanisms, distribution priority and manual intervention at a data level; however, the existing solutions suffer from the following disadvantages: (1) The system lacks automatic fusing/bypassing and quick recovery capability when the system encounters problems such as abnormal check-in service, wrong rule template, diagnosis and diagnosis of a consulting room, fluctuation of data quality or equipment failure on site, and is easy to cause long-time unresponsiveness or abnormal sequencing; (2) The unified dynamic decision is lacking, namely the unified health assessment and weight scheduling mechanism is lacking among multisource rules (sign-in, display number strategy, priority crowd, fairness penalty and consulting room load balance), and the self-adaption is difficult under an abnormal scene; (3) Factors such as punishment of number passing, manual queue insertion, priority of review and the like are coupled in a complex manner, and the fairness and efficiency are difficult to ensure at the same time in the prior practice; (4) The shadow evaluation and backtracking are absent, namely the shadow contrast evaluation and audit are absent in the strategy switching and sequencing result, and the effectiveness and compliance controllability of the recovery strategy are difficult to prove; In view of the above, the invention provides an outpatient queue AI decision bus method and system based on dynamic fusing and self-healing. Disclosure of Invention Based on the technical problems in the background technology, the invention provides an outpatient queue AI decision bus method and an AI decision bus system based on dynamic fusing and self-healing. The invention provides an outpatient queue AI decision bus method based on dynamic fusing and self-healing, which comprises the following steps: S1, data acquisition and preprocessing, namely acquiring state, sorting, priority, load and sign-in rule data from a patient table and an associated table; S2, health degree monitoring, namely constructing a health probe for a check-in switch, a display number strategy, load statistics and automatic diagnosis marking, and generating health degree H and fault indexes; s3, fusing judgment, namely maintaining a fuse state machine aiming at each rule/data source, and triggering a degradation strategy when abnormality occurs; S4, selecting a strategy set, namely selecting a main strategy set according to the current health state, wherein the main strategy set comprises a display number, a priority constraint, fairness punishment and load balancing weight; s5, AI multi-factor scoring, namely scoring and sorting candidate patients, outputting a queue sequence and a display number, and simultaneously reserving a shadow strategy for parallel scoring; S6, shadow evaluation, namely counting the difference of a shadow strategy and a main strategy on key indexes, and providing evidence for semi-open heuristic/restoration closure; S7, half-open trial and self-healing, namely switching the small-proportion flow to an alternative strategy when the shadow strategy counted in S6 shows stability, and observing the stability and then recovering to be closed; S8, audit record and visualization, namely recording factor constitution, weight, fusing state and switching reason of each decision, and supporting traceability and supervision; s9, online learning and weight self-adaption, namely automatically adjusting the weight and the threshold value of each factor based on the historical evaluation data; S10, outputting the existing queue, and accessing the existing interface and page in a compatible mode to ensure backward compatibility. Preferably, in the step S2, the health degree H is calculated by using the formula H=1-alpha. FAILRATEI-beta. TimeoutRatei-gamma. MISSINGRATEI-delta. ConflictRatei; Wherein FAILRATEI is error rate, timeoutRatei is timeout rate, MISSINGRATEI is critical field missing rate, conflictRatei is rule conflict rate, and α, β, γ and δ are configurable weights. P