Individual matching and team complementary intelligent formation decision system for rescue tasks

Abstract

This invention relates to the technical field of intelligent grouping decision-making systems for individual matching and team complementarity in rescue missions. Specifically, it discloses such a system, aiming to address the problems of static and fixed job models, insufficient individual adaptation, and inadequate team collaboration in traditional rescue grouping. The system includes a job competency model construction unit, a dynamic member profile management unit, a two-layer optimized grouping decision-making unit, a task assignment support unit, and a personalized development plan generation unit. By constructing a self-evolving multi-dimensional job competency model, combined with a stable member competency profile, and employing a two-layer optimization mechanism of individual-job matching and team-task complementarity, it generates highly adaptable grouping schemes and links with teaching resource knowledge graphs to push personalized development paths. This invention enables closed-loop management of scientific rescue team grouping, risk warning, and continuous talent development, improving emergency response efficiency and collaboration levels.

Description

Technical Field

[0001] This invention belongs to the field of artificial intelligence and emergency rescue collaborative decision-making technology, specifically involving an intelligent grouping decision-making system for individual matching and team complementarity for rescue missions. Background Technology

[0002] With the increasing frequency of sudden public safety incidents worldwide, efficient coordination and precise team formation in emergency rescue missions have become core elements for improving rescue success rates. Modern rescue systems rely on the scientific evaluation of rescue team members' capabilities and a deep analysis of rescue mission requirements, aiming to maximize rescue effectiveness through rational personnel allocation. In complex and ever-changing extreme environments, the success or failure of rescue operations often depends on the precise matching of professional skills among team members and the comprehensive application of interdisciplinary knowledge, which places high demands on the scientific rigor and real-time nature of rescue decision-making.

[0003] Intelligent team formation decision-making technology based on individual characteristic profiling is a key direction for achieving scientific deployment. This technology quantifies multi-dimensional data such as the professional skills, physical fitness, and psychological state of rescue personnel to construct matching models for different rescue positions. Based on this, it searches for the optimal team combination scheme from a talent pool using specific mathematical models or algorithms. Its basic principle is to use a data-driven approach to achieve precise matching between task requirements and human resources, ensuring that rescue missions can be carried out in an orderly and efficient manner under various constraints.

[0004] Most existing job competency evaluation models are based on pre-set static indicator systems, making it difficult to dynamically evolve and spontaneously optimize based on successful cases generated in rescue practice. This leads to a gradual disconnect between model parameters and actual job requirements. Traditional team formation schemes often focus on decentralized analysis of individual job matching, lacking in-depth optimization of overall team competency coverage, role complementarity among members, and avoidance of key weaknesses, making it difficult to form synergy at the overall operational level. Current rescue management processes generally suffer from the problem of isolated talent assessment and subsequent training, failing to automatically link teaching resources and generate feedback loops for competency gaps identified in team formation decisions.

[0005] Therefore, an intelligent grouping decision-making system for individual matching and team complementarity is desired for rescue missions. (Invention Content) The purpose of this invention is to provide an intelligent grouping decision-making system for individual matching and team complementarity for rescue missions, which can solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows: An intelligent grouping decision-making system for individual matching and team complementarity in rescue missions includes a job competency model building unit, a dynamic team member profile management unit, a two-tiered optimized grouping decision-making unit, a mission assignment support unit, and a personalized development plan generation unit, among which: The job competency model construction unit is used to establish multi-dimensional competency requirement vectors for various rescue positions, and dynamically adjust the competency requirement vectors based on the competency characteristics of high-performing team members in historical missions, so as to achieve self-evolution of the job model. The dynamic profile management unit for team members is used to collect and integrate multi-source data on each rescue team member's professional skills, psychological qualities, physical condition and combat performance, etc., to construct their steady-state capability profile, and continuously update it to reflect their capability evolution trajectory. The dual-layer optimization grouping decision unit is used to calculate the individual matching degree between available team members and target positions after receiving specific rescue mission requirements, and identify key capability gaps. Then, with the required job combination as a constraint, it comprehensively optimizes the overall capability coverage of the team, the complementarity of roles among members, and the degree of avoidance of key weaknesses, and generates one or more optimal or suboptimal team grouping schemes, along with an analysis of the advantages of each scheme and potential risk warnings. The task dispatch support unit is used to automatically call the recommended grouping scheme and its evaluation basis generated by the two-layer optimized grouping decision unit when a new rescue mission is triggered, so as to provide the commander with scientific and efficient decision support. The personalized development plan generation unit is used to automatically assemble a personalized development roadmap, which includes theoretical learning content, simulation training projects and classic case studies, based on the ability gaps identified in the team member profiles and in conjunction with the teaching resource knowledge graph. The roadmap is then pushed to the corresponding team member and their manager.

[0007] Preferably, the multi-dimensional capability requirement vector constructed by the job capability model construction unit covers the dimensions of professional skills, psychological qualities, and physical fitness, with each dimension having several sub-indicators to accurately characterize the capability requirements of a specific rescue job.

[0008] Furthermore, the job competency model construction unit continuously monitors the results of historical task execution, selects team members who perform well in specific positions, and performs cluster analysis on their competency profiles. When the clustering results remain stable and better than the benchmark level within a predetermined time, the cluster center is automatically extracted as a new competency requirement reference, and the original job competency requirement vector is corrected.

[0009] Furthermore, the steady-state capability profile constructed by the team member dynamic profile management unit consists of long-term stable capability characteristics, which is different from temporary state data. It is dynamically updated through regular evaluation and task feedback mechanisms to ensure that the profile always reflects the team member's true capability level.

[0010] Preferably, when the dual-layer optimized grouping decision unit performs individual-job matching degree calculation, it uses a vector similarity algorithm to compare the steady-state capability profile of team members with the job capability requirement vector, and outputs a matching degree score and a specific capability gap list.

[0011] Furthermore, when performing team-task complementarity optimization, the dual-layer optimization grouping decision unit sets the required job combination for the task as a hard constraint condition, and takes maximizing the overall capability coverage of the team, minimizing the number of key capability shortcomings, and enhancing the complementarity of roles and capabilities among members as multi-objective optimization criteria. It performs a global search in the pool of available team members to generate a set of grouping schemes that meet the preset optimization threshold.

[0012] Furthermore, each team grouping scheme generated by the dual-layer optimized grouping decision unit is accompanied by a structured explanatory document, which elaborates on the advantages of the grouping in terms of task adaptability, risk control capabilities and collaborative potential, and points out potential blind spots in capabilities or hidden dangers in collaboration.

[0013] Preferably, after receiving the task type information, the task dispatch support unit automatically parses the job configuration list required for the task and triggers the two-layer optimization grouping decision unit to perform grouping calculations, and finally presents the sorted recommended squad list along with its evaluation criteria to the commander terminal.

[0014] Furthermore, the teaching resource knowledge graph associated with the personalized development plan generation unit includes a curriculum system, training scenario library, case database, and capability improvement path templates, which can intelligently match the most suitable learning and training content according to the type and severity of capability gaps.

[0015] Furthermore, the personalized development roadmap generated by the personalized development plan generation unit has the functions of setting phased goals and tracking progress, supporting managers to conduct closed-loop supervision and effect evaluation of the team members' ability improvement process.

[0016] Compared with the prior art, the present invention has the following beneficial effects: 1. The intelligent grouping decision system for individual matching and team complementarity for rescue missions provided by this invention constructs a self-evolving job competency model, enabling job requirements to be dynamically adjusted according to the best performance in rescue practice, thus avoiding the problem of traditional static models being out of touch with actual needs; by introducing a two-layer optimization mechanism of "individual matching-team complementarity", the scientific nature and synergy of rescue grouping are improved. 2. By establishing a closed-loop link from capability assessment to personalized training, this invention enables the system not only to be used for mission decision-making, but also to drive the continuous growth of team members and improve the quality of talent reserves and response capabilities of the entire rescue system.

[0017] 3. This system provides solid technical support for efficient, accurate, and intelligent grouping in complex emergency scenarios, and has outstanding practical value and prospects for promotion. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the overall technical solution architecture of the present invention; Figure 2 This is a schematic diagram of the core principle framework of individual matching and team complementarity in the two-layer optimized grouping decision unit of this invention; Figure 3 This is a flowchart of the self-evolution logic of the job model based on historical performance feedback in the job competency model construction unit of this invention. Figure 4 This is a schematic diagram of the multi-level interaction relationship and data flow of the task dispatch support unit in a specific rescue task triggering scenario in this invention; Figure 5 This is a schematic diagram of the closed-loop improvement principle based on capability gap identification and knowledge graph linkage in the personalized development plan generation unit of this invention. Detailed Implementation

[0019] Example 1: Please refer to the appendix Figure 1 To be continued Figure 5 To make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to specific embodiments.

[0020] An intelligent grouping decision-making system for individual matching and team complementarity in rescue missions includes a job competency model construction unit, a team member dynamic profile management unit, a two-layer optimized grouping decision-making unit, a mission dispatch support unit, and a personalized development plan generation unit. The job competency model construction unit establishes multi-dimensional competency requirement vectors for various rescue positions and dynamically adjusts these vectors based on the competency characteristics of high-performing team members in historical missions to achieve self-evolution of the job model. The team member dynamic profile management unit collects and integrates multi-source data on each rescue team member's professional skills, psychological qualities, physical condition, and combat performance to construct their steady-state competency profile, which is continuously updated to reflect their competency evolution trajectory. The two-layer optimized grouping decision-making unit, upon receiving a specific rescue mission requirement, first calculates the available team members and... The system assesses the individual matching degree between target positions and identifies key capability gaps. Then, using the required job combination as a constraint, it comprehensively optimizes the team's overall capability coverage, role complementarity among members, and the degree of avoidance of key weaknesses, generating one or more optimal or suboptimal team grouping schemes, along with an analysis of the advantages and potential risks of each scheme. The task assignment support unit automatically invokes the recommended grouping scheme and its evaluation basis generated by the two-layer optimization grouping decision unit when a new rescue mission is triggered, providing commanders with scientific and efficient decision support. The personalized development plan generation unit automatically assembles a personalized development roadmap, including theoretical learning content, simulation training projects, and classic case studies, based on the capability gaps identified in the team member profiles and in conjunction with the teaching resource knowledge graph, and pushes the roadmap to the corresponding team members and their managers.

[0021] Furthermore, the job competency model construction unit is physically deployed on a highly reliable cloud management server, which is equipped with a large-scale structured database for storing a standard job terminology database for the rescue field. The job competency model construction unit includes a job dimension definition subunit, a quantitative indicator configuration subunit, and a self-evolving calculation engine. The job dimension definition subunit is configured to perform attribute modeling for at least 30 typical rescue jobs, such as "on-site first aider," "transfer commander," "demolition and rescue worker," and "communication support worker." Each attribute model covers at least three dimensions: professional skills, psychological qualities, and physical fitness. Under the professional skills dimension, it is further subdivided into at least five secondary indicators, such as theoretical knowledge reserves, proficiency in operating tools, and cross-disciplinary collaboration ability. Under the psychological qualities dimension, it is further subdivided into at least four secondary indicators, such as stress tolerance, decision-making decisiveness, and emotional stability. Under the physical fitness dimension, it is further subdivided into at least three secondary indicators, such as cardiorespiratory endurance, explosive power, and weight-bearing walking ability. The quantitative indicator configuration subunit is used to set a normalized standard score between 0 and 1 for each of the above secondary indicators, forming a multi-dimensional capability requirement vector for a specific position. This multi-dimensional capability requirement vector serves as the initial benchmark for the position requirements.

[0022] More specifically, the self-evolving computation engine in the job competency model construction unit is configured with a dynamic adjustment logic based on feedback control. This self-evolving computation engine continuously monitors the execution results of historical rescue missions and automatically extracts mission reports rated as "excellent" or "high-performance" from the mission management system. The engine calls historical data from the team member dynamic profile management unit via an interface to filter team member samples holding key positions in specific high-performance missions and performs cluster analysis on the competency profiles of these team members during mission execution. The self-evolving computation engine employs an improved density clustering algorithm to identify cluster centers of high-performance team member competency characteristics in a multi-dimensional vector space. When the system detects that a cluster center for a certain position remains stable for more than three consecutive time evaluation periods, and the comprehensive evaluation index corresponding to this cluster center is better than the currently preset job competency requirement vector (e.g., exceeding the current benchmark value by more than 15%), the self-evolving computation engine automatically generates a weight adjustment instruction. This instruction shifts the current job competency requirement vector towards the high-performance cluster center, with the adjustment step dynamically adjusted based on the richness of historical successful samples. Through this mechanism, the job competency model is no longer a static description of clauses, but a dynamic benchmark that can automatically evolve as the "best practices" characteristics shown in actual combat are demonstrated.

[0023] The team member dynamic profile management unit acts as the data foundation of the system, connecting to multiple heterogeneous data source interfaces. This unit includes a multi-source data acquisition subunit, a steady-state capability modeling subunit, and a profile dynamic update subunit. The multi-source data acquisition subunit is configured to interface with the daily training management system, psychological monitoring terminal, physical fitness assessment equipment, and combat feedback recorder. The types of data collected include, but are not limited to: team members' operational precision values ​​in simulated training, heart rate variability data in psychological stress tests, various competitive results in physical fitness qualifying competitions, and qualitative and quantitative scores given by commanders in past missions. The steady-state capability modeling subunit uses a long short-term memory neural network model to denoise the collected time-series data, eliminating abnormal fluctuations caused by accidental factors (such as exceptional performance or unexpected mistakes in a single mission), and extracting a "steady-state capability layer" profile that represents the team member's long-term, stable qualities. This profile is stored in a high-frequency cache database in the form of a multi-dimensional vector, with each component corresponding to a specific capability index score. The dynamic profile update subunit is configured to incrementally update the steady-state capability layer vector of the team members according to a preset time step (such as once a month) or triggered by a specific event (such as after completing a major rescue mission), so as to ensure that the timeliness of the profile can keep pace with the team members' actual skill evolution trajectory.

[0024] The dual-layer optimized grouping decision unit is the core computing node of this system, integrating complex operations research optimization algorithms. This unit includes an individual matching degree calculation module, a team complementarity optimization module, and a grouping report generation module. Upon receiving the job configuration requirements for a specific rescue mission, the individual matching degree calculation module first executes the first-layer matching logic: it calls the steady-state capability profile vectors of all members in the currently available team member pool and performs vector inner product calculation or Euclidean distance mapping with the requirement vectors of each target job required by the mission. This calculation process not only outputs a scalar score representing the overall matching degree (e.g., 0.85 represents a high fit), but also accurately identifies the key capability gaps of the team member relative to the target job through component subtraction operations. For example, if a team member has a matching degree of 0.9 for the "demolition and rescue worker" job, but their "explosive power" component is lower than the corresponding value in the job requirement vector, the system marks "insufficient explosive power" as a key gap.

[0025] Furthermore, the team complementarity optimization module executes a second layer of optimization logic. This module sets the required job combinations for the task (e.g., 1 commander, 2 first responders, 2 rescuers) as a hard constraint on the optimization problem. Under the premise of meeting the hard quantity requirements of the positions, this module is configured as a multi-objective optimization solver. The first objective is to maximize the overall capability coverage of the team, that is, to ensure that the overlap of team members' capabilities in their professional fields is minimized, while maximizing the breadth of coverage of the skills required for the task. The second objective is to minimize the number of critical weaknesses, that is, to ensure that there are no common weaknesses within the team that could affect the success or failure of the task through personnel combinations. The third objective is to enhance the complementarity of roles and capabilities among members, for example, combining an experienced but slightly less physically fit veteran team member with a physically fit but less experienced new team member. The solver performs a global search among tens of thousands of possible combinations using a heuristic search algorithm or a genetic algorithm, ultimately selecting several grouping schemes that meet the preset optimization threshold.

[0026] The team formation report generation module is responsible for transforming the calculated abstract data into decision support information readable by commanders. For each generated team formation plan, the module automatically calls upon a built-in risk assessment knowledge base to generate a structured explanatory document. The document details the advantages of the plan in addressing specific mission risks (e.g., "ample medical rescue capabilities"), while objectively pointing out potential risks (e.g., "team member coordination is still in the adjustment period"). This transparent information presentation significantly reduces the cognitive load on commanders under extreme pressure.

[0027] The task dispatch support unit is deployed on the command center's large screen and mobile handheld terminals. This unit features automatic task type parsing. When a new rescue task is triggered, the system automatically extracts the corresponding job configuration list from the task template library based on the input disaster briefing (including key information such as disaster type, scale of damage, and environmental parameters). For example, for a "high-rise building fire rescue" task, the system automatically sets the required proportions of communication, firefighting, search and rescue, and medical positions. Subsequently, this unit automatically triggers the two-layer optimized team formation decision unit to perform real-time calculations and presents the sorted recommended team list in the form of a visual matrix. Commanders can instantly view detailed capability profiles, historical task performance, and responsibilities within the current team by clicking on personnel avatars in the matrix. The system also supports manual "fine-tuning intervention" by commanders, allowing them to replace individual team members based on unforeseen circumstances. The system will then recalculate the overall team effectiveness score and risk coefficient in real time, achieving scientific decision-making through human-machine collaboration.

[0028] The personalized development plan generation unit constructs a closed loop from talent assessment to precise training. This unit integrates a rescue teaching resource knowledge graph. This knowledge graph is stored in the form of a graph database, where nodes represent different teaching resources, such as "hydraulic clamp operation video tutorials," "high-voltage electric shock simulation training scripts," and "complex weather navigation case study reports," etc.; edges represent the supporting relationships and difficulty levels between resources and capability indicators. The personalized development plan generation unit is configured such that when the dual-layer optimization grouping decision unit identifies a team member's capability gap, it automatically uses the capability gap label as a search seed to search for related paths in the knowledge graph. The system not only selects theoretical learning materials directly corresponding to the gap, but also automatically assembles a personalized development roadmap containing three stages: "theoretical self-study - simulated practice - case study," based on the team member's rank and historical training records. For example, for a team member experiencing setbacks in "psychological stability," the roadmap will prioritize online courses on psychological stress intervention, then link to appointment slots in psychological simulation training rooms, and push successful rescue cases in similar extreme environments for their study. The generated roadmap will be automatically pushed to the team members' mobile app and synchronized to their direct supervisor's management backend, supporting real-time progress tracking and periodic assessments to ensure that every identified weakness can be addressed in a targeted manner.

[0029] Furthermore, to ensure the system's distributed operation capabilities and data security, the intelligent grouping decision-making system for individual matching and team complementarity for rescue missions also includes a distributed data synchronization subsystem. This subsystem employs a consensus mechanism based on blockchain technology to ensure that training data and mission performance evaluations generated locally by rescue teams at all levels are immutable and have timestamp traceability during the uploading process to the centralized profile database. Simultaneously, when performing grouping decision calculations, the system uses hardware acceleration cards (such as multi-core GPUs or FPGA arrays) to handle high-concurrency vector similarity operations and multi-constraint combination optimization operations, ensuring that a complete grouping suggestion can be provided within 30 seconds during the golden rescue period after a disaster. The system's network architecture uses a dedicated encrypted link, coupled with a two-factor authentication mechanism, to strictly prevent the leakage of sensitive personnel files and mission deployment information.

[0030] In this embodiment, the job competency model building unit also introduces a time-weighted decay mechanism during self-evolution. For successful cases that are far removed from the current time, their contribution weight to the job model correction decreases exponentially with time. This mechanism takes into account the evolution of rescue technologies, equipment, and mission environments over time. For example, an efficient "search and rescue job" 10 years ago might have focused on sensory search, while today it must include the ability to rapidly process UAV thermal imaging data. Through time-weighted decay, the system can prioritize absorbing recent technological development trends, ensuring that the job competency model always remains at the forefront of the rescue industry.

[0031] In the construction of the dynamic profile management unit for team members, the system also features a specially designed cold start processing logic. For newly recruited team members lacking practical experience, the system maps them to a preset "young eagle profile template" based on their pre-recruitment professional background, initial physical fitness test scores, and personality test results using a correlation analysis algorithm. As the team member participates in more simulation exercises, the system gradually introduces real-world data to replace the template until a complete, personalized, stable capability profile is constructed. This smooth transition mechanism solves the problem of lacking reference data in new personnel grouping decisions.

[0032] The dual-layer optimized grouping decision unit also supports multi-dimensional weight customization during execution. Commanders can manually adjust the emphasis on "matching degree priority" or "complementarity priority" based on the urgency or specificity of the task. For example, when executing extremely high-risk individual assault missions, commanders can increase the weight of "individual job matching degree" to ensure that each position is filled by the strongest expert in that field; while when executing long-term comprehensive post-disaster reconstruction missions, the weight of "role complementarity and team stability" can be increased to ensure the team's mental health and operational endurance during long-term collaboration. This flexible parameter adjustment capability allows the system to adapt to various extreme application scenarios.

[0033] The personalized development plan generation unit also integrates an intelligent scheduling algorithm when pushing out roadmaps. This algorithm can automatically compare the current duty roster of the brigade with the individual training plans of team members, and automatically arrange learning and practical tasks using fragmented time or planned training days without affecting combat readiness. The system has an automatic reminder function; when a team member fails to complete the learning progress on time, it will automatically send a progress warning to the team member and the manager, realizing the digitalization and closed-loop management of training.

[0034] In summary, the intelligent grouping decision-making system for individual matching and team complementarity for rescue missions provided in this embodiment achieves synchronized alignment between standards and practices by constructing a self-evolving job model; it transforms isolated individual capabilities into synergistic team combat effectiveness through a two-layer optimization algorithm; and it realizes continuous value-added of human resources through an automated link from profiling to planning. The application of this system will enhance the response speed and operational effectiveness of rescue organizations in complex and uncertain environments.

[0035] Example 2: Based on the intelligent grouping decision system for individual matching and team complementarity for rescue missions described in Example 1, this example provides a variant of the system architecture based on edge computing and mobile deployment. This architecture is specifically designed for the self-organizing grouping needs in extreme disaster environments such as road closures, power outages, and network outages.

[0036] In this alternative system architecture, the job competency model building unit, the team member dynamic profile management unit, the two-layer optimized grouping decision-making unit, the task assignment support unit, and the personalized development plan generation unit are highly integrated into a portable, ruggedized computing terminal. This ruggedized computing terminal is waterproof, dustproof, and shockproof, and has a built-in high-capacity low-temperature battery array, enabling it to work continuously for more than 48 hours in extremely cold or humid environments.

[0037] In this embodiment, the job competency model building unit is configured to have localized, simplified computational capabilities. When a portable terminal enters a disaster area lacking network coverage, it automatically loads the job competency requirement vector last synchronized from the cloud as its initial state. Although it cannot obtain high-performance cases from a global database in real time, this unit can perform micro-evolutionary calculations based on localized task performance within the current frontline command's jurisdiction. For example, at a specific mudslide site, if a rescue team demonstrates extremely high demolition efficiency, the system will instantly analyze the current profile characteristics of that team members and make localized temporary adjustments to the job requirement vector under the same terrain conditions, achieving on-site adaptation of the algorithm to specific disaster scenarios.

[0038] In this architecture, the team member dynamic profile management unit communicates in real-time with the smart sensing nodes worn by team members via short-range low-power wireless communication protocols (such as ZigBee, Bluetooth Low Energy, or proprietary narrowband radio signals). These sensing nodes are built into rescue helmets, smart exoskeletons, and physiological monitoring wristbands, enabling them to acquire dynamic parameters such as team members' blood oxygen saturation, body temperature, cadence, and remaining equipment battery power in real time. The team member dynamic profile management unit then weights and fuses these real-time dynamic parameters with pre-stored steady-state capability profiles. For example, if a team member has a high-level "steady-state physical fitness score," but real-time monitoring shows that their heart rate recovery period is prolonged due to overexertion, the system will automatically lower their "physical fitness availability" coefficient at the current moment. This profile correction based on real-time physiological feedback is crucial for ensuring the safety of team formation decisions in the field.

[0039] The dual-layer optimization grouping decision unit employs a hierarchical dimensionality reduction computation strategy when executed on the local terminal. Considering the computational limitations of portable devices, during team-task complementarity optimization, the system first performs preliminary screening based on the distance between team members' current locations and task points, eliminating members who cannot reach the task within the specified time, thus narrowing the search space. Subsequently, the system uses a lightweight multi-objective particle swarm optimization algorithm for grouping search. This algorithm can generate at least three optimal solutions within 5 seconds when processing a pool of dozens of candidates. Each solution also undergoes complementarity verification to ensure the team's collaborative capabilities under constrained conditions.

[0040] The task assignment support unit presents information via a high-brightness touchscreen display built into the terminal. In outdoor environments with strong sunlight, the screen automatically switches to a high-contrast black-and-white display mode and displays grouping suggestions using clear and concise graphical icons (such as radar charts and organizational trees). Commanders can confirm personnel and issue task assignment instructions via gestures. Due to the harsh communication environment in disaster areas, the unit also integrates a Mesh self-organizing network routing module. Confirmed grouping instructions are transmitted step-by-step to the mobile terminals of all selected team members through multi-hop relays, achieving "penetrating" instruction delivery.

[0041] The personalized development plan generation unit automatically switches to "offline knowledge base" mode when offline. The portable terminal has a built-in, streamlined knowledge graph of teaching resources, as well as commonly used rescue manuals, standard operating procedures (SOPs), and multimedia instructional videos. When the system detects a team member's skill gap due to temporary assignment to an unfamiliar support role during team formation decisions, the unit immediately extracts core skill points from the offline database and sends "pre-battle crash course key points" to the team member's terminal via offline push functionality. This immediate knowledge replenishment effectively mitigates the coordination risks associated with temporary team formations.

[0042] Furthermore, to address the coordination issues of multiple portable terminals on-site, the system in this embodiment also includes a master-slave backup synchronization protocol. When multiple terminals are set up at the on-site command center, the system automatically elects one as the master control node for global optimization calculations, with the remaining terminals serving as redundant backups. If the master control node suffers physical damage, the backup nodes will take over all tasks within milliseconds, ensuring the continuity of the grouping decision-making process. Once the network returns to normal, all locally generated decision records and profile update data will be automatically uploaded back to the cloud master database, completing the global profile database.

[0043] This edge computing-based system architecture variant complements the cloud architecture in Example 1. The cloud architecture focuses on deep learning with massive amounts of data and long-term job model evolution, while the edge architecture emphasizes real-time perception, rapid response, and reliable operation in extreme environments. Through periodic data alignment, they construct a comprehensive intelligent grouping and decision-making system covering the cloud, network, and edge.

[0044] In practical applications of this embodiment, the system demonstrated extremely high flexibility. During a simulated post-earthquake urban search and rescue exercise, after the communication base station was damaged, the portable terminal quickly took over the management of 30 team members on site. By collecting physiological parameters in real time, the system accurately identified two team members who were severely dehydrated due to continuous work and immediately replaced them as reserve team members in the next round of regrouping recommendations. Simultaneously, based on the complex internal structure of the damaged buildings, the system automatically adjusted the capability requirement vector for the "searcher" position, increasing the weight of "three-dimensional spatial positioning capability."

[0045] Example 3: Building upon the aforementioned examples, this example provides a system expansion solution for large-scale cross-regional rescue collaboration. This solution focuses on addressing how to achieve cross-agency individual matching and team complementarity when multiple rescue organizations and different functional departments are operating jointly.

[0046] In this embodiment, the job competency model construction unit is configured to have a semantic conversion interface. Since different rescue forces (such as fire and rescue teams, the Red Cross, and professional volunteer organizations) have different naming and evaluation standards for positions, the semantic conversion interface uses natural language processing technology to map the original job descriptions of each organization to the standard job terminology of this invention. For example, it performs equivalence checks between an organization's "Level 1 First Aid Responder" and this system's "Senior On-Site First Aid Responder," and re-aligns the quantitative scores based on the historical assessment data of each organization. This unified "competency language" is a prerequisite for large-scale cross-regional collaboration.

[0047] The team member dynamic profile management unit incorporates distributed digital identity (DID) technology in this solution. Each team member's competency profile is encapsulated as a digital proof with privacy-preserving attributes. During cross-organizational calls, the system only retrieves competency components relevant to the current task (such as "demolition competency score"), without exposing sensitive personal information of team members (such as age, medical history, etc.). This data interaction mechanism based on the principle of least privilege addresses data security and privacy concerns in cross-organizational collaboration. Simultaneously, the system supports federated learning of heterogeneous data, meaning that without aggregating the original data from various organizations, the self-evolving computational engine for job roles can be jointly trained and optimized, enabling the model to learn from a broader range of case samples.

[0048] In this embodiment, the two-tier optimized grouping decision unit is upgraded to a multi-level coordination architecture. The first level is intra-organizational grouping, where each rescue organization generates preliminary job matching plans in its respective standby pool. The second level is global balancing, where the two-tier optimized grouping decision unit optimizes personnel replacement and capability complementarity among multiple rescue organizations using a game theory-based resource allocation algorithm. For example, when organization A lacks sufficient diving search and rescue capabilities, while organization B has redundant manpower in this area, the system automatically generates cross-organizational mixed grouping suggestions. The optimization objective not only includes capability coverage but also additionally considers "cross-organizational collaboration costs," prioritizing pairing with members from organizations that have a history of joint exercises.

[0049] The task dispatch support unit features multi-level command chain adaptation capabilities. It can automatically switch command views based on the scale of the task. In large-scale disaster scenarios, commanders can view a cloud map of troop deployment and capability distribution across the entire area; within specific combat units, it provides detailed personnel matching information. The system supports cross-platform command delivery; regardless of the type of walkie-talkie or mobile device used by team members, commands can be formatted and delivered through a unified gateway protocol.

[0050] In a cross-institutional scenario, the personalized development plan generation unit evolves into an "educational resource sharing alliance." The system can link online courses and training site information from different rescue organizations. When a firefighter needs to improve their rescue capabilities for a specific type of chemical fire, but internal fire department resources are temporarily scarce, the system can, guided by a knowledge graph, recommend and request access to open teaching resources from nearby chemical companies or private emergency academies. This cross-sectoral training model improves the efficiency of emergency rescue talent cultivation across society.

[0051] Furthermore, to ensure fairness and scientific rigor in decision-making during large-scale collaboration, this embodiment also introduces a grouping effectiveness retrospective evaluation subsystem. After each rescue mission, this subsystem automatically compares the deviation between the "system-recommended grouping" and the "actual performance." Using reinforcement learning algorithms, the system continuously adjusts the parameter weights in the individual matching degree calculation. If data shows that a certain type of complementary combination performs poorly in multiple practical scenarios, the system automatically lowers the recommendation priority of such combinations. This closed-loop learning process enables the system to demonstrate increasingly higher levels of intelligence when dealing with complex, large-scale rescue missions.

[0052] The cross-regional collaborative architecture of this embodiment has been empirically demonstrated in a flood and drought emergency drill in a certain province. In this drill, the system successfully achieved unified grouping decisions for rescue forces from five cities and three different types of personnel across the province. When faced with a sudden breach-sealing task, the system dispatched the most suitable expert team and operational forces from other locations within 3 minutes, and scientifically mixed them based on the complementarity of the members, verifying the superior efficiency of this invention in handling the grouping of complex socialized rescue forces.

[0053] Those skilled in the art should understand that although the functions of each unit and its subunits have been described in detail in the foregoing embodiments, the physical implementation of these functions is not limited to specific hardware devices. The units can be integrated into a single server cluster or deployed across multiple layers such as cloud, edge, and endpoint. The logic of data interaction, similarity calculation, and optimized search can be implemented using various equivalent mathematical methods or computational models without departing from the spirit of this invention.

[0054] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0055] To further enhance the system's resilience and robustness, a redundant logic verification module can be configured in all embodiments of this system. This module monitors the operating status of each computing unit in real time. When it detects that an optimization algorithm is trapped in a local optimum or that a computation timeout occurs due to abnormal input data, the module automatically triggers a preset baseline rule engine. Based on the hard requirements in the rescue manual (such as "each group must be equipped with at least one licensed doctor"), the baseline rule engine quickly generates an initial grouping scheme that meets the bottom-line safety requirements. This serves as a backup for the intelligent scheme, ensuring that the rescue grouping decision-making process will not be interrupted under any extreme hardware or software failure.

[0056] Furthermore, both the job competency requirement vector and the team member steady-state competency profile in this invention support dynamic dimensional expansion. With future upgrades to rescue equipment (such as exoskeletons, quadruped robots, and brain-computer interface communication devices), managers only need to define new competency indicators in the management interface, and the system will automatically include those indicators in similarity calculations and complementarity optimization, without requiring refactoring of the core code. This open system architecture ensures that this invention has an extremely long technological lifespan and broad scenario adaptability.

[0057] At the software engineering level, this system is built using a microservice architecture, with units such as job modeling, profile management, and group optimization running as independent microservices. These microservices communicate asynchronously via lightweight message queues, significantly improving the system's concurrent processing capabilities. For the massive amounts of stored team member profile data, sharded database technology is used for management, supporting real-time retrieval at the terabyte level. The system's front-end interface adopts a responsive design, adapting to various terminal devices such as tablets, smartphones, and command screens, ensuring optimal information presentation in different rescue scenarios.

[0058] Finally, this system strictly adheres to data anonymization and auditing requirements during data processing. All sensitive operations involving team member identities are recorded in an indelible system audit log. When conducting large-scale data statistical analysis, the system employs differential privacy technology to add appropriate noise to the raw data, ensuring both the accuracy of the statistical results and preventing the inference of specific team member privacy information from statistical data. This rigorous approach to data ethics makes the system more compliant and acceptable in practical deployment.

[0059] In summary, this invention addresses the core pain points of traditional rescue team building, such as rigid models, poor coordination, and a disconnect between assessment and training, through a complete technical architecture encompassing "job modeling, individual profiling, two-layer optimization, dispatch support, and development closed loop." The implementation of this system will propel my country's emergency rescue management towards digitalization, intelligence, and scientific rigor, yielding both social security value and economic benefits.

[0060] The vector calculations and logical decisions involved in the above embodiments are described in more detail below: In the process of calculating the individual-job fit, the system performs a multi-criteria decision analysis. First, the system extracts each capability component from the team member's profile and subtracts it from the corresponding scalar value in the job requirement vector to obtain a deviation vector. Then, the system weights this deviation vector, with weight coefficients dynamically assigned by the job model building unit based on the importance of the indicator in the current task. For example, in extremely dangerous environments, the weight coefficient for "psychological stability" will be set to the highest. Next, the system calculates the norm of the weighted deviation vector and converts it into a fit score between 0 and 1. The closer the score is to 1, the more suitable the team member is for the job requirements on that specific indicator. Finally, the system integrates the fit scores of all indicators and obtains the team member's total fit score for the job through a weighted average.

[0061] In the team complementarity optimization process, the system defines "complementarity" as the non-overlapping nature of team members' skill weaknesses. Specifically, the system constructs a team capability matrix, where each row represents a member's capability vector. The system uses logical operations to identify "weaknesses" in each row that are below a benchmark. One optimization goal is to transform the row combinations so that each column of the matrix contains at least one high-scoring item (i.e., a strength), effectively "covering" any individual's weakness at the team level. Simultaneously, the system assesses "role complementarity" by calculating the orthogonality between member capability vectors. The closer the vectors are to orthogonality, the less overlap in capabilities among members, and the greater the potential for synergy. The system identifies the optimal balance between this orthogonality and the sum of the matching degrees to determine the final recommended solution.

[0062] This computational process, which is entirely based on textual logical description, has an underlying logic that corresponds perfectly to the functions of the system units. It can guide those skilled in the art to implement the core functions of the present invention through logic programming without relying on specific mathematical formulas.

[0063] In summary, the intelligent grouping decision-making system for rescue missions, which combines individual matching and team complementarity, provides by constructing a self-evolving job competency model, allows job requirements to be dynamically adjusted based on optimal performance in rescue practice, effectively avoiding the problem of traditional static models being out of touch with actual needs. By introducing a two-tiered optimization mechanism of "individual matching-team complementarity," it achieves a leap from single-job adaptation to maximizing overall team efficiency, significantly improving the scientific rigor and coordination of rescue grouping. By establishing a closed-loop link from competency assessment to personalized training, the system can not only be used for mission decision-making but also drive the continuous growth of team members, improving the quality of talent reserves and responsiveness of the entire rescue system. This system provides solid technical support for efficient, accurate, and intelligent grouping in complex emergency scenarios, possessing outstanding practical value and promising prospects for widespread application.

[0064] In any of the above embodiments, the connection relationships between the modules, units, or sub-units include wired connections and / or wireless connections. The wired connections include, but are not limited to, physical media such as optical fibers, twisted pairs, and coaxial cables. The wireless connections include, but are not limited to, Wi-Fi, Bluetooth, cellular networks (such as 4G, 5G, and 6G), satellite communication, and microwave communication. The storage device can be a hard disk drive (HDD), a solid-state drive (SSD), flash memory, cloud storage, or any other form of non-transitory storage medium.

[0065] The terms "first," "second," etc., used in this invention are only used to distinguish different components or functions and do not represent any specific order or degree of importance. Where there is no conflict, the technical features in the various embodiments can be combined arbitrarily to form new implementation schemes. These combined schemes also fall within the protection scope of this invention.

[0066] Finally, it should be noted that the embodiments provided in this specification are merely preferred embodiments of the present invention. Those skilled in the art can make various equivalent transformations according to actual application needs without departing from the core concept of the present invention. For example, more sensor types can be added to enrich the data source for team member profiles, or more constraint variables (such as geographical distance between team members, real-time wear status of equipment, etc.) can be introduced into the grouping optimization algorithm. All such extensions based on the logical framework of the present invention should be considered as part of the present invention.

Claims

1. An intelligent grouping decision-making system for individual matching and team complementarity in rescue missions, characterized in that: include: The job competency model construction unit is configured to establish a multi-dimensional competency requirement vector consisting of professional skills, psychological qualities, and physical fitness for various rescue positions, and to dynamically offset and correct the multi-dimensional competency requirement vector based on the competency characteristics of high-performing team members in historical missions, so as to achieve the self-evolution of the job model. The team member dynamic profile management unit is configured to collect and integrate multi-source data of rescue team members in actual combat and training through heterogeneous data source interfaces, use neural network models to extract steady-state capability layer profiles that represent the long-term stable qualities of team members, and perform incremental updates based on the team members' capability evolution trajectory. The two-layer optimized grouping decision unit is configured to, upon receiving specific rescue mission job requirements, first calculate the individual matching degree between available team members and target positions and identify key capability gaps, and then, using the job combination required for the mission as a hard constraint, comprehensively optimize the team's overall capability coverage, the complementarity of roles among members, and the degree of avoidance of key shortcomings, and generate a team grouping scheme with risk assessment instructions. The task dispatch support unit is configured to parse the job configuration list when a new rescue task is triggered, and call the recommended grouping scheme and its evaluation basis generated by the two-layer optimized grouping decision unit to provide decision support information to the commander terminal; The personalized development plan generation unit is configured to automatically assemble and push a personalized development roadmap that includes theoretical learning, simulation training, and case studies, based on the capability gaps identified by the team member dynamic profile management unit and in conjunction with a knowledge graph containing teaching resources and their relationship with capability indicators.

2. The intelligent grouping decision-making system for individual matching and team complementarity for rescue missions as described in claim 1, characterized in that, The job competency model construction unit includes: a job dimension definition subunit, which is configured to perform attribute modeling for rescue jobs. The attribute modeling covers professional skills dimension, psychological quality dimension and physical fitness dimension, and sets multiple secondary indicators under each dimension. Professional skills include theoretical knowledge reserves, proficiency in operating tools, and cross-disciplinary collaboration ability; The dimensions of psychological qualities include stress tolerance, decision-making decisiveness, and emotional stability. The physical fitness dimensions include cardiorespiratory endurance, explosive power, and weight-bearing walking ability; The quantitative indicator configuration subunit is configured to set normalized standard scores for the secondary indicators, forming an initial multidimensional capability requirement vector for a specific rescue position. The self-evolving computing engine is configured as a dynamic adjustment component based on feedback control. The self-evolving computing engine continuously monitors the execution results of historical rescue missions and extracts high-performance mission reports. It calls historical data in the team member dynamic profile management unit through an interface to perform vector space clustering analysis on the capability profiles of high-performance team members during mission execution.

3. The intelligent grouping decision-making system for individual matching and team complementarity for rescue missions as described in claim 2, characterized in that, When performing dynamic offset correction, the self-evolving computing engine adopts the following logic: the self-evolving computing engine uses a density clustering algorithm to identify the clustering centers of the ability characteristics of high-performing team members in a multi-dimensional vector space; When the cluster center of a job position is detected to remain spatially stable within a preset number of time evaluation periods, and the comprehensive evaluation index corresponding to the cluster center is better than the current preset job capability requirement vector, the self-evolutionary calculation engine automatically generates a weight adjustment instruction. The weight adjustment instruction drives the current job competency requirement vector to shift and correct towards the high-performing team member cluster center, and the correction step size is dynamically adjusted according to the number of historical successful samples. The self-evolutionary computing engine also introduces a time-weighted decay mechanism, which is configured to reduce the contribution weight of historical cases that are far removed from the current time point to the correction of the job model, so that the multi-dimensional capability requirement vector can preferentially absorb the recent evolution trend of rescue technology.

4. The intelligent grouping decision-making system for individual matching and team complementarity for rescue missions as described in claim 1, characterized in that, The team member dynamic profile management unit includes a multi-source data acquisition subunit, which is configured to interface with the training management system, psychological monitoring terminal, physical fitness assessment equipment and combat feedback recorder to collect data including the team member's operational accuracy in simulated training, heart rate variability data in psychological stress tests, competitive performance in physical fitness assessments and evaluation scores in tasks. The steady-state capability modeling subunit is configured to use a long short-term memory neural network model to denoise the collected time-series data, remove abnormal data fluctuations caused by single task environment or accidental factors, extract steady-state capability components that can represent the long-term and stable quality of team members, and store them in a high-frequency cache database in the form of multi-dimensional vectors. The portrait dynamic update subunit is configured to incrementally update the steady-state capability layer vector of the team members according to a preset time step or triggered by the completion of a major rescue mission. The cold start processing module is configured to map newly recruited team members who lack practical experience to a preset "Young Eagle" profile template based on their pre-recruitment professional background, initial physical fitness test scores, and personality test results using an association analysis algorithm. The "Young Eagle" profile template is then gradually replaced as the actual test data increases.

5. The intelligent grouping decision-making system for individual matching and team complementarity for rescue missions according to claim 1, characterized in that, The dual-layer optimized grouping decision unit includes: an individual matching degree calculation module, configured to extract the steady-state capability layer profile vector of members in the available team member pool, and perform vector inner product calculation or Euclidean distance mapping between it and the multi-dimensional capability requirement vector of each target position required by the task, outputting a matching degree score representing the degree of fit between the individual and the position, and identifying the specific capability gap items of the member relative to the target position through component subtraction operation; The team complementarity optimization module is configured as a multi-objective optimization solver. Under the premise of meeting the hard constraints of task and job combination, the team complementarity optimization module takes maximizing the overall capability coverage of the team, minimizing the number of key capability shortcomings, and enhancing the complementarity of roles and capabilities among members as multi-objective optimization criteria. It uses heuristic search algorithms or genetic algorithms to perform a global search in the candidate team member combinations and select a set of team grouping schemes that meet the preset optimization threshold. The team formation report generation module is configured to call the built-in risk assessment knowledge base to write a structured explanatory document for each generated team formation scheme. The explanatory document includes an analysis of the scheme's advantages in dealing with task risks, potential blind spots in capabilities, and warnings of potential collaboration risks.

6. The intelligent grouping decision-making system for individual matching and team complementarity for rescue missions according to claim 5, characterized in that, When evaluating the complementarity of roles and abilities among members, the team complementarity optimization module executes the following logic: The team complementarity optimization module constructs a team capability matrix, where each row represents the capability vector of a team member; The multi-objective optimization solver transforms the row combination so that each column of the team capability matrix has at least one capability component that is higher than the preset benchmark, thereby achieving coverage of individual capability shortcomings at the team level. The team complementarity optimization module evaluates role complementarity by calculating the degree of orthogonality between the ability vectors of different members. The more orthogonal the vectors between members are, the smaller the overlap of their abilities and the greater their synergy potential. The team complementarity optimization module also supports multi-dimensional weight customization, which can be configured to manually adjust the weight ratio between the matching score and the role complementarity according to the urgency or special nature of the task, so as to achieve grouping emphasis adjustment for different extreme application scenarios.

7. The intelligent grouping decision-making system for individual matching and team complementarity for rescue missions according to claim 1, characterized in that, The task dispatch support unit includes: a task type automatic parsing subunit, configured to extract the corresponding job configuration list from the built-in task template library based on the disaster type, disaster scale and environmental parameters in the input disaster briefing; The visual interactive terminal is configured to present the recommended grouping scheme output by the two-layer optimized grouping decision unit in the form of a visual matrix, and allows commanders to view detailed capability profiles, historical task performance and job responsibilities by clicking on the personnel avatars in the matrix. The fine-tuning intervention component is configured to receive personnel fine-tuning instructions from the commander for the grouping plan, and after fine-tuning, trigger the two-layer optimized grouping decision unit in real time to recalculate the overall team performance score and risk coefficient after personnel replacement, so as to realize human-machine collaborative decision correction.

8. The intelligent grouping decision-making system for individual matching and team complementarity for rescue missions according to claim 1, characterized in that, The personalized development plan generation unit includes: a rescue teaching resource knowledge graph, which is stored in the form of a graph database. Its nodes represent teaching resources including video tutorials, simulation training scripts, and case study reports, and its edges represent the supporting relationship and difficulty level between teaching resources and ability indicators. The personalized route assembly module is configured to use the team member capability gap labels identified by the dual-layer optimized grouping decision unit as search seeds to perform associated path searches in the rescue teaching resource knowledge graph, and automatically assemble a personalized development roadmap that includes a theoretical self-study stage, a simulated practice stage, and a case study stage based on the team member's rank and historical training records. The intelligent scheduling subunit is configured to compare the brigade's duty schedule with the individual training plans of team members. Without affecting combat readiness, it can automatically arrange learning and practical tasks using fragmented time and has a progress warning function, supporting managers to conduct closed-loop supervision and effect evaluation of the capability improvement process.

9. The intelligent grouping decision-making system for individual matching and team complementarity for rescue missions according to any one of claims 1 to 8, characterized in that, The system also includes a portable ruggedized computing terminal, which integrates the job competency model building unit, the team member dynamic profile management unit, the two-layer optimized grouping decision unit, the task dispatch support unit, and the personalized development plan generation unit. The portable ruggedized computing terminal is equipped with a low-temperature battery array and has a tri-proof structure, which is used to provide self-organizing grouping decision in extreme disaster environment such as circuit failure, power failure, and network failure. The job competency model building unit in the portable ruggedized computing terminal has localized simplified computing capabilities. It is configured to automatically load the job competency requirement vector that was last synchronized in the cloud when there is a lack of network coverage, and to perform localized job model correction based on the current localized task performance. The team member dynamic profile management unit in the portable ruggedized computing terminal connects to the intelligent sensing node worn by the team member via a short-range wireless communication protocol, acquires the team member's physiological dynamic parameters in real time, and performs weighted fusion of the physiological dynamic parameters with the pre-stored steady-state capability layer profile to adjust the team member's availability coefficient at the current moment.

10. The intelligent grouping decision-making system for individual matching and team complementarity for rescue missions according to any one of claims 1 to 8, characterized in that, The system also includes a distributed data synchronization subsystem, which adopts a consensus mechanism based on blockchain technology to ensure that the training data and mission performance evaluations generated locally by rescue teams at all levels are immutable and have timestamp traceability during the process of uploading to the centralized profile database. Hardware acceleration components, including multi-core processors, graphics processing arrays, or field-programmable gate arrays, are configured to process vector similarity operations and multi-constraint combination optimization operations in the two-layer optimization grouping decision unit; The semantic conversion interface is configured to use natural language processing technology to map the original job descriptions of different rescue organizations to the standard job vocabulary of the job competency model building unit, thereby achieving cross-organizational competency evaluation standard alignment. The distributed digital identity module is configured to encapsulate team members’ capability profiles into digital proofs with privacy-preserving attributes, and only exposes capability components related to the current task when called across organizations. The redundancy logic verification module is configured to monitor the operating status of each unit in real time. When an optimization calculation anomaly or timeout is detected, the built-in baseline rule engine is automatically triggered to quickly generate an initial backup grouping scheme based on the rescue safety baseline requirements.

Images