A multi-role cooperation-oriented management procurement process dynamic evaluation system

Abstract

The application provides a management procurement process dynamic evaluation system and method for multi-role cooperation, relates to the technical field of management decision auxiliary and semantic big data processing, and solves the problem that the traditional time sequence process is difficult to quantify the cooperation value. Firstly, the procurement intention evolution tree is constructed, the operation characteristic embedding is carried out by using the semantic increment model of non-Euclidean space, the evolution track of the procurement demand from the fuzziness to the accuracy is captured in real time in the dynamic topology, the problem that the traditional time sequence process is difficult to quantify the cooperation value is solved, the multi-dimensional cooperation friction recognition mechanism is adopted, the semantic alignment constraint is introduced, the interaction behavior is decoupled into the effective deterministic contribution and the invalid shock loss, the deep attribution of the cooperation efficiency is realized, the dynamic credit image is used for the elastic process reconstruction, the role ability and the environmental interference variable are distinguished, and the process optimization strategy and the dynamic intervention instruction with the execution guiding force are output. The application effectively improves the substantial explainability of the evaluation result, and enhances the response speed and the scientificity of the procurement decision in the complex cooperation environment.

Description

Technical Field

[0001] This invention relates to the field of management decision support and semantic big data processing technology, specifically a dynamic evaluation system for management procurement processes oriented towards multi-role collaboration. Background Technology

[0002] As enterprises deepen their digital transformation, procurement management has evolved from simple material purchases to complex multi-role collaborative activities involving supply chain coordination, cash flow control, and compliance audits. Existing enterprise resource planning systems, supplier management systems, and office automation systems typically use workflow engines to realize the electronic flow of procurement business. In these systems, the procurement process is defined as a linear sequence of a series of preset nodes: "Initiate application - Department approval - Financial review - Procurement execution - Acceptance and payment".

[0003] To improve procurement efficiency, companies often need to evaluate and manage the operation of the procurement process. Current technologies primarily rely on statistical analysis of structured log data for evaluation, including:

[0004] Timeliness analysis: Calculate the total time taken from the initiation to the end of the process, or the dwell time of each approval node at a specific stage, in order to assess personnel efficiency.

[0005] Compliance check: Check whether the document fields are complete, whether attachments are uploaded, and whether the budget is exceeded, in order to determine whether the process is compliant.

[0006] Results-oriented evaluation: Analyze the difference between the final purchase price and historical prices to assess the cost savings rate.

[0007] However, the aforementioned existing technologies have the following significant technical shortcomings and management blind spots when dealing with the complex and ever-changing collaborative scenarios of modern enterprises:

[0008] First, there is a lack of in-depth measurement of the "value of interactive content" (semantic blind spot). Existing technologies mainly focus on the flow of process status (i.e., "whether it is completed" and "when it is completed"), while ignoring the substantive contributions of roles to the information content during the flow. Approving personnel only pursue timeliness indicators and perform "instant approval" without conducting substantive review of procurement needs; or procurement staff frequently modify documents, which, although showing high activity, is actually due to rework caused by insufficient communication in the early stage. The existing evaluation system based on timestamps and status codes cannot distinguish between "efficient progress" and "perfunctory work", nor can it identify "effective iteration" and "ineffective idle work", resulting in distorted evaluation results that cannot truly reflect the quality of collaboration.

[0009] Second, the evaluation and process control are disconnected, lacking dynamic adaptability (static rigidity). Most existing procurement processes use rigid templates with static configurations. Evaluation is often conducted offline after the fact, and the evaluation results cannot be fed back to the currently running process instance in real time. This means that regardless of whether the cooperation between the collaborating parties has a very high or very low historical level of cooperation, the system forces the same approval path. For high-trust collaboration chains, redundant approval nodes cause inefficiency. For collaboration nodes with potential risks, the system lacks the ability to dynamically add verification mechanisms, making it difficult to achieve flexible process control that is "personalized for each individual".

[0010] Third, it ignores the implicit collaboration costs and non-linear iterative characteristics. Actual procurement collaboration is often non-linear, involving a large amount of repeated communication, rejection and modification, and comparison of multiple versions. Existing technologies usually simplify the process into a directed acyclic graph, which makes it difficult to capture the "cyclical oscillation" caused by unclear requirements or the "branching divergence" caused by information asymmetry. The system lacks quantitative means for this kind of "collaboration friction", which makes it impossible for managers to locate the root cause of process obstruction (whether it is a human problem or an information transmission mechanism problem).

[0011] In summary, existing technologies require an evaluation system that can penetrate the surface of processes, deeply understand changes in the semantics of interactions, and dynamically reconstruct process paths in real time based on collaboration quality. Summary of the Invention

[0012] Technical problems to be solved

[0013] To address the shortcomings of existing technologies, this invention provides a dynamic evaluation system for multi-role collaborative management and procurement processes, solving the following problems:

[0014] 1. Solved the pseudo-activity problem of "difficulty in quantifying contribution" during collaboration:

[0015] Traditional systems cannot distinguish between "ineffective diligence" (frequent modifications but no substantial progress) and "low-frequency high-value decisions." By using semantic atomic capture and effective incremental calculation, this invention eliminates repetitive, rollback, and redundant interference information, truly reflecting the role's substantial driving force for achieving procurement goals, and solving the problem that evaluation indicators are superficial and cannot penetrate the essence of business.

[0016] 2. Resolved the mismatch between rigid business processes and dynamic collaboration requirements:

[0017] The existing process, no matter how well the collaborating parties cooperate, must go through the same fixed nodes, which leads to the institutional loss of efficiency of "familiar people and high-reputation links". By introducing a flexible process restructuring mechanism, the flow protocol is dynamically adjusted by using the "collaboration efficiency profile" generated by real-time evaluation. This achieves "process dehydration" of high-reputation paths and "logic reinforcement" of high-risk paths, solving the "rigidity" problem that the system cannot evolve itself according to the actual collaboration quality.

[0018] 3. Solved the problem of "ambiguous localization of the root cause of obstruction" in complex collaborative networks:

[0019] When the procurement process slows down, managers often only see "which link is stuck" but do not know whether it is due to poor information quality provided by the preceding node or the node's own weak processing capacity. By constructing a procurement intent evolution tree, the system can trace the "divergence" and "convergence" process of information. If a node frequently causes subsequent semantic regression, it can be identified as the root cause of friction. This solves the problems of ambiguous responsibility determination and inaccurate bottleneck diagnosis in complex multi-role networks.

[0020] 4. It resolved the monitoring blind spot caused by the disconnect between informal communication and formal processes:

[0021] Many core decisions are made through communication tools or offline, and formal documents cannot reflect the true collaboration process, resulting in incomplete evaluation data. By using cross-modal alignment technology to map unstructured communication with structured documents, implicit collaboration logic can be captured, solving the problems of fragmented evaluation data and lack of compliance monitoring caused by "information silos".

[0022] 5. Solved the problems of "information entropy increase" and uncertainty management in the decision-making process:

[0023] Procurement requirements often become increasingly chaotic during the transfer process, leading to a deviation between the final delivery and the original intention. The system monitors the evolution of procurement intentions in real time. Once it detects that the intention has deviated from the root node (initial requirement), it immediately triggers an alert and outputs optimization instructions, thus solving the problem of the continuous decline in the "fidelity" of procurement requirements in large organizations during the transfer of multiple roles.

[0024] Technical solution

[0025] To achieve the above objectives, the present invention provides the following technical solution: a dynamic evaluation system for management and procurement processes oriented towards multi-role collaboration, comprising:

[0026] Sp1: The full-link interactive semantic atomization capture module is used to monitor the operation behavior of each collaborative role in the procurement process, and parse each document editing, approval opinion submission, parameter modification and instant messaging reply into semantic atomic operation units with timestamps and role identifiers;

[0027] Sp2: Procurement Intent Evolution Tree Construction Module, used to construct a procurement intent evolution tree that reflects the information state changes of the procurement object in the collaboration chain, with the initial procurement demand as the root node and the semantic atomic operation unit as the growth node. The evolution tree records the complete topological trajectory of procurement information from fuzzy to precise.

[0028] Sp3: Semantic Increment and Collaborative Friction Calculation Engine, used to traverse the procurement intent evolution tree, calculate the effective semantic increment value of each role node by comparing the information differences between parent and child nodes, and identify collaborative friction nodes that cause semantic regression or generate cyclic branches.

[0029] SP4: Dynamic credit weighting and flexible process reconfiguration module, used to generate a collaborative performance profile of each role in real time based on historical semantic increment values ​​and the frequency of collaborative friction nodes, and dynamically adjust the current procurement process flow protocol according to the profile. The adjustment of the flow protocol includes automatic merging of approval nodes, skipping or adding mandatory verification steps.

[0030] Preferably, the calculation logic for the semantic effective increment value in Sp3 is as follows: the system identifies the modification of the information of the previous node by the subsequent node as a positive increment, invalid redundancy, or negative interference; when the modification of the subsequent node makes the procurement specification parameters clearer and the compliance clauses more complete, it is determined to be a positive increment; when the modification of the subsequent node is revoked or repeatedly modified in subsequent steps, it is determined to be invalid redundancy; the system accumulates the positive increment and deducts the invalid redundancy to obtain the actual contribution value of the role in the current process, which serves as the core basis for dynamic evaluation.

[0031] Preferably, the process elastic reconfiguration module in Sp4 has a dynamic approval threshold drift function: the system maintains a dynamic trust library based on collaboration efficiency profiles; when it detects that the semantic increment of historical collaboration between the initiator and the approver is continuously positive and the friction is extremely low, the system automatically raises the exemption threshold of the link to achieve automatic acceleration of low-risk processes; when it detects that the input information of a specific role causes subsequent nodes to frequently trigger correction operations, the system automatically lowers the operation authority of the role and forcibly inserts a preprocessing verification plugin in its submission stage.

[0032] Preferably, the procurement intent evolution tree construction module in Sp2 includes branch pruning logic: during the process, when the system identifies that the discussion or modification of a branch is not ultimately merged into the main procurement intent, the branch is marked as a trial branch; the system counts the participation of each role in the trial branch and uses the participation as an indicator to evaluate the role's risk investigation ability, rather than a simple efficiency indicator, so as to distinguish between ineffective work and necessary risk exploration.

[0033] Preferably, the system also includes an implicit dependency mining module: used to analyze the high-frequency semantic atomic operation associations of different roles under non-process-defined paths; when it is found that the decision of role A is highly dependent on the informal input of role B, the system automatically establishes implicit coupling weights between A and B in the evaluation model; when evaluating the performance of A, the system combines the response quality of B for weighted calculation, thereby quantifying the impact of informal collaboration on formal processes.

[0034] Preferably, the semantic atomization capture module in Sp1 has a cross-modal alignment function: mapping unstructured communication text to structured ERP document fields; when a commitment to price or delivery date appears in the communication text but is not updated synchronously in the ERP document, the system generates a consistency warning signal and uses this signal as a negative evaluation parameter for the degree of collaborative integrity.

[0035] Preferably, the dynamic evaluation method for the management procurement process executed by the system includes the following steps:

[0036] Sp101: Initialize the procurement intent evolution tree and parse the procurement requisition form into a root information node; Sp102: Capture the derived operations of each role on the root information node in real time, generate child nodes and attach them to the evolution tree; Sp103: Compare the differences between each parent-child node pair, calculate the semantic increment of the current operation, and determine whether the operation promotes process convergence or causes divergence; Sp104: Update the collaborative performance profile of the role based on the accumulated semantic increment; Sp105: Based on the updated profile, adjust the triggering conditions or approval paths of subsequent process nodes in real time and output an evaluation report.

[0037] Beneficial effects

[0038] This invention provides a dynamic evaluation system for management and procurement processes oriented towards multi-role collaboration. It offers the following advantages:

[0039] 1. It achieves precise quantification of collaborative value, eliminating the evaluation pitfall of "ineffective work":

[0040] Unlike existing technologies that only count node dwell time, this invention uses "semantic increment" calculation to accurately identify which roles are substantially advancing the project and which roles are merely performing redundant operations. Through pruning and analysis of "trial and error branches," the system can scientifically distinguish between "necessary discussions to avoid risks" and "inefficient rework due to insufficient capabilities," making the evaluation results more reflective of employees' professional level rather than just response speed.

[0041] 2. A self-healing and adaptive mechanism for the process has been built, significantly improving overall collaboration efficiency:

[0042] This invention breaks the rigid limitations of traditional procurement processes. Based on a real-time updated "collaboration efficiency profile," the system can automatically implement a "green channel" strategy (automatically reducing nodes) for high-trust, high-value collaboration paths, and a "strengthened compliance" strategy (increasing verification) for high-friction, low-incremental-value links. Approval limits and permissions are no longer static and rigid configurations, but dynamically fluctuate according to the credit and performance of both collaborating parties. This flexible restructuring mechanism can shorten the circulation cycle of conventional low-risk procurement by more than 30%.

[0043] 3. It has enabled in-depth diagnosis of collaboration bottlenecks, moving from "phenomenon localization" to "root cause tracing":

[0044] By using "procurement intent evolution tree" and "semantic fallback" analysis, managers can intuitively see at which stage information becomes ambiguous or contradictory, thus accurately identifying whether it is a misunderstanding by a specific role or a technical defect in the initial requirement description, uncovering implicit dependencies under non-process paths, revealing the real information flow trajectory between departments, helping enterprises to discover and solidify those informal but efficient collaboration patterns, and optimizing the organization's communication topology.

[0045] 4. Enhanced information certainty and compliance throughout the entire procurement lifecycle:

[0046] Cross-modal alignment (comparison of ERP documents and instant messaging records) can intercept the risk of "inconsistency between verbal promises and system records" in real time. The tamper-proof value ledger: combined with semantic evolution records, the system not only records "who did what" but also the intentional logic of "why it was done". In the face of post-audit or supplier disputes, it can provide a more convincing causal logic chain than traditional operation logs. Attached Figure Description

[0047] Figure 1 This is a cloud diagram illustrating the system composition of the present invention;

[0048] Figure 2 This is a system architecture diagram of the present invention;

[0049] Figure 3 This is a flowchart of the process of the present invention. Detailed Implementation

[0050] 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. Specific Implementation Example 1:

[0052] like Figure 1-3 As shown, a dynamic evaluation system for management and procurement processes oriented towards multi-role collaboration includes:

[0053] Sp1: The full-link interactive semantic atomization capture module is used to monitor the operation behavior of each collaborative role in the procurement process, and parse each document editing, approval opinion submission, parameter modification and instant messaging reply into semantic atomic operation units with timestamps and role identifiers;

[0054] Sp2: Procurement Intent Evolution Tree Construction Module, used to construct a procurement intent evolution tree that reflects the information state changes of the procurement object in the collaboration chain, with the initial procurement demand as the root node and the semantic atomic operation unit as the growth node. The evolution tree records the complete topological trajectory of procurement information from fuzzy to precise.

[0055] Sp3: Semantic Increment and Collaborative Friction Calculation Engine, used to traverse the procurement intent evolution tree, calculate the effective semantic increment value of each role node by comparing the information differences between parent and child nodes, and identify collaborative friction nodes that cause semantic regression or generate cyclic branches.

[0056] SP4: Dynamic credit weighting and flexible process reconfiguration module, used to generate a collaborative performance profile of each role in real time based on historical semantic increment values ​​and the frequency of collaborative friction nodes, and dynamically adjust the current procurement process flow protocol according to the profile. The adjustment of the flow protocol includes automatic merging of approval nodes, skipping or adding mandatory verification steps.

[0057] Sp1. Link Interaction Semantic Atomization Capture:

[0058] This step is executed by the collaborative metadata real-time awareness module, which captures the entire lifecycle of procurement operations through event listening probes pre-installed in the user interface of the business system and protocol parsing interfaces in the backend of the communication service.

[0059] The specific implementation logic is as follows: The system monitors every operational event in the procurement business process in real time. When document editing, status changes, or information sending are detected, the system parses and reconstructs the unstructured behavioral data into standardized semantic atomic units. These semantic atomic units logically contain the following core information fields:

[0060] End-to-end tracing identifier: used to bind discrete operational behaviors to specific procurement transaction instances, ensuring the contextual integrity of data flow.

[0061] Behavior subject identifier: Records the role identity and permission attributes of the person performing the operation.

[0062] Time dimension attribute: Records the precise timestamp of the operation, used to build subsequent time-series logic.

[0063] Semantic content change: Text difference comparison technology is used to extract the amount of change in information content before and after the operation, and natural language processing model is used to transform unstructured text into high-dimensional feature vectors.

[0064] Operation Intent Label: Based on a preset behavior classification library, the system automatically determines whether the operation is promotional, rejection, cancellation, or consultation-type, and labels the corresponding intent attribute.

[0065] Sp2. Construction of the procurement intention evolution tree:

[0066] This step is performed by the role-based collaborative digital twin building module, which abandons the traditional linear linked list storage method and instead adopts a directed acyclic graph data structure to dynamically map the evolution of procurement intentions.

[0067] The specific implementation logic is as follows:

[0068] Root node establishment: The system uses the initialized purchase request form as the root information node of the evolution tree, which represents the original baseline state of the purchase intention.

[0069] Node growth mechanism: Whenever Sp1 captures a semantic atomic unit with substantial semantic changes, the system generates a new derived node based on the current operation context. The derived node is connected to the parent node through a directed edge, which represents the direction of information flow and processing time.

[0070] Branching and merging logic:

[0071] Parallel branch generation: When the system detects that different roles have performed mutually exclusive modification operations based on the same parent node, the system automatically generates parallel sibling node branches on the evolution tree and establishes conflict association markers between the branches to represent the state of disagreement in the collaboration process.

[0072] Path convergence and merging: When subsequent operations reach a consensus on the divergent content or execute the approval action, the scattered branch paths converge to a new merge node, representing the achievement of collaborative consensus.

[0073] Invalid branch marking: For modification attempts that are not ultimately merged into the main path or for drafts that are rejected, the system marks them as pruned to account for trial and error costs in subsequent analysis and collaboration processes.

[0074] Sp3. Semantic Increment and Cooperative Friction Calculation:

[0075] This step is performed by the semantic increment and cooperative friction computation engine, which is based on graph theory algorithms and information theory principles to perform real-time traversal and quantitative analysis of the evolution tree constructed by Sp2.

[0076] The specific implementation logic is as follows:

[0077] Semantic effective increment evaluation: The system evaluates semantic increment by calculating the information entropy difference between the parent node and the child node.

[0078] Deterministic gain: The system analyzes the fuzzy language density and parameter precision in the node text. When a child node significantly reduces the uncertainty of the description compared to its parent node (i.e., information entropy is reduced), the system determines that the operation has produced a positive semantic effective increment.

[0079] Compliance gain: The system matches the changed content with a pre-built compliance risk terminology library. When risk coverage clauses are added, the effective incremental value is further accumulated.

[0080] Redundancy deduction: When the content of a child node is only a format adjustment or a character change without any substantial meaning, the system will determine that its increment is zero or an extremely low value.

[0081] Evaluation of Cooperative Friction Coefficient: Systematically analyze the topological characteristics of the evolutionary tree.

[0082] Rollback friction: The system detects whether there is a semantic regression path pointing to the content of the previous version, that is, the content of the subsequent node actually restores the state of the previous node. Such paths are defined as invalid rollbacks and are included in the friction coefficient.

[0083] Oscillation Friction: The system counts the number of cyclical branches and interaction rounds generated in the same decision-making process. The more cyclical rounds, the greater the resistance to reaching consensus among roles. Based on this, the system calculates a high coefficient of collaborative friction.

[0084] SP4. Dynamic Credit Weighting and Flexible Process Restructuring:

[0085] This step is executed by the Dynamic Credit Weighting and Flexible Process Restructuring module, which transforms the calculation results of Sp3 into real-time control instructions for the business process engine, enabling adaptive adjustment of the process.

[0086] The specific implementation logic is as follows:

[0087] Collaboration efficiency profile update: Based on historically accumulated semantic effective increments and collaboration friction coefficients, the system updates the dynamic trust index of each participating role in real time. This index reflects the role's professionalism and cooperation in collaboration.

[0088] Flexible process refactoring strategy:

[0089] Path optimization (fast track mechanism): When the dynamic trust index of all roles involved in the current collaboration link is higher than the preset high reputation threshold, and the risk attribute of the current procurement transaction is lower than the dynamic exemption threshold, the system automatically triggers a process jump instruction. This instruction drives the workflow engine to automatically execute the passage operation of non-critical approval nodes, thereby achieving physical acceleration of the process.

[0090] Risk intervention (slow lane mechanism): When a sudden increase in the collaboration friction coefficient of the current link is detected, and a continuous negative value of the effective semantic increment (i.e. semantic drift) is detected, the system automatically triggers the process circuit breaker or reinforcement instruction. This instruction drives the workflow engine to lock the submission permission of the current node and dynamically inject mandatory manual review nodes and compliance verification tasks until the system detects a new semantic increment that meets the stability requirements before the restriction can be lifted and the process flow can be restored. Specific Implementation Example 2:

[0092] like Figure 1-3 As shown, based on the content of the above-described specific embodiment one, the following content is further disclosed:

[0093] Details of building a collaborative digital twin:

[0094] This step uses graph neural network embedding technology to achieve a precise mapping of physical collaborative behavior to digital space. The system establishes dynamic entity nodes for each participating role in the virtual space.

[0095] Feature vector extraction: The system continuously extracts the dynamic behavioral features of role nodes, including the consistency of response latency, the professional dimension of semantic output, and the boundary frequency of cross-departmental interactions, and encapsulates them into multi-dimensional feature vectors.

[0096] Dynamic topology monitoring: The system constructs a weighted directed graph with roles as vertices and interactive semantic flows as edges. By monitoring the stability of the graph structure in real time, the system can identify topological holes caused by the absence of key roles or the breakage of response chains. Once a hole is formed, the digital twin model will immediately issue a synchronous warning and simulate the delay impact of the hole on subsequent process nodes.

[0097] Adaptive weight allocation mechanism:

[0098] The evaluation model has an environment-aware weight adjustment function. The system not only relies on preset static indicators, but also emphasizes the dynamic alignment between evaluation standards and business objectives.

[0099] Dimensional weight matrix: The system pre-sets four core evaluation dimensions: compliance, efficiency, cost, and stability.

[0100] Fuzzy comprehensive evaluation application: Based on the priority characteristics of the current procurement task, the system uses fuzzy set theory to linearly recombine the weights of each dimension. In emergency scenarios, the system automatically increases the weight of time responsiveness and reduces the weight sensitivity of cost saving rate to ensure that the evaluation results are in line with the current business focus.

[0101] Collaboration bottleneck cause-and-effect diagnosis logic:

[0102] The system uses counterfactual reasoning techniques to penetrate superficial data and pinpoint the true root causes of obstacles to collaboration.

[0103] Mediation effect analysis: The system calculates the mediation weight of a specific node in the information flow. When a node frequently intercepts information and causes subsequent nodes to generate a large number of redundant queries, the system will determine that the node has a high mediation negative effect.

[0104] Bottleneck tracing: The system automatically traces the contextual factors that cause a node to stall, distinguishing whether it is due to the limited processing capacity of the role or the excessive information entropy input from the preceding links, which prevents it from making a decision.

[0105] Asymmetric information recognition mechanism:

[0106] This module aims to eliminate hidden barriers in the collaboration process.

[0107] Information symmetry assessment: The system compares the overlap of information shared among the three parties—the procurement requester, the executor, and the supplier—in real time.

[0108] Behavioral deviation detection: By analyzing the discrepancies between the intent expressed in the communication logs and the actual parameters entered into the business system, the system identifies any instances of information hiding or false feedback. Information transparency will be directly used as a negative correction factor for the credit profile.

[0109] Interactive interference suppression control:

[0110] To avoid excessive evaluation leading to a burden on collaboration, the system introduces a task load awareness mechanism.

[0111] Command noise reduction logic: The system monitors the task stack depth of each role's terminal in real time. When it is determined that a role is in an overloaded state, the system automatically performs delayed push or information aggregation processing on non-urgent process optimization commands to ensure that the evaluation intervention behavior does not turn into new collaborative interference.

[0112] Collaborative trajectory consensus storage technology:

[0113] Distributed ledger technology is used to ensure the authority and immutability of the evaluation results.

[0114] Decision point hash storage: The system generates an encrypted digest of the semantic state, evaluation score, and intervention instructions automatically triggered by the system for each key decision point.

[0115] Accountability Matrix Tracing: The evidence records form a collaborative accountability chain across time and space, providing legally valid digital evidence for subsequent process compliance audits.

[0116] IoT sensing and consistency comparison with physical records:

[0117] The system extends the evaluation perspective to the physical world, achieving a virtual-real integrated evaluation.

[0118] Real-time logistics status access: By integrating geolocation data and sensing interfaces, the system can obtain the physical displacement status of materials in real time.

[0119] Logical verification: The system automatically compares the physical location information with the inbound nodes within the system. When there is a temporal and spatial logical conflict between the physical displacement and the document flow, the system will automatically lower the performance evaluation level of the relevant operation role.

[0120] Dynamic evaluation methods and computer storage media logic:

[0121] Method execution flow: The system first initializes the task ontology library to determine the baseline behavioral characteristics; then, it continuously monitors and updates the digital twin model synchronously; it uses game equilibrium algorithms and causal diagnosis to determine the deviation value and its root cause; finally, it performs multi-path deduction in virtual space, selects and executes the optimized instructions that can maximize the overall collaborative increment.

[0122] Media operation mechanism: The instruction set in the storage medium drives the hardware processor to execute all the above-mentioned logical judgments, data calculations and process scheduling behaviors, ensuring the automated operation of the nursing procedure. Specific Implementation Example 3:

[0124] like Figure 1-3 As shown, based on the content of the above specific embodiments, the following content is further disclosed:

[0125] The core working principle of this system lies in transforming traditional static procurement management into a closed-loop process of "dynamic perception, logical deduction, and real-time control." It automatically adjusts the management intensity by analyzing the information quality in the collaboration process in real time, thereby achieving a balance between efficiency and risk.

[0126] The following is a detailed explanation of how the system works:

[0127] Phase 1. Full capture and standardization of collaborative information:

[0128] The system first builds a full-link perception network. During the procurement process, when various roles perform operations, the system does not only record "who completed the task", but also uses underlying plugins and interfaces to capture every subtle change in the interaction content.

[0129] Atomized parsing: The system parses all interactions (modifying contract terms, discussing delivery dates in instant messaging, approving documents) into semantic atomic units. These units contain the operating subject, timestamp, specific information changes, and the operator's attitude.

[0130] Cross-modal alignment: The system can associate unstructured dialogue records with structured document fields in real time. When a buyer commits to a new price during communication, the system will automatically map it to the version of the inquiry form in the system, ensuring that all collaborative information is logically consistent and traceable.

[0131] Phase Two: Dynamic Construction of the Procurement Intent Evolution Tree

[0132] This is the "brain" of the system. The system uses graph database technology to draw a procurement intention evolution tree in real time as the process progresses, starting from the initial procurement needs.

[0133] Path growth: Every effective modification or confirmation will cause the evolution tree to grow new branches. By comparing the differences between parent nodes and child nodes, the system can clearly see whether the procurement intention has become more precise or more chaotic due to poor communication.

[0134] Disagreement and consensus positioning: When different departments have different opinions on the same procurement parameters, the evolution tree will generate parallel branches. The system automatically identifies the points of conflict in the collaboration process by monitoring the duration of these branches. When all parties reach a consensus and merge the branches, the system determines that the collaboration has entered the next stable stage.

[0135] Phase Three. Quantitative Determination of Collaborative Value and Friction Loss:

[0136] The system uses an evolutionary tree topology and runs a semantic incremental evaluation algorithm to assess the value of each participant's contribution.

[0137] Semantic effective increment determination: The system analyzes the deterministic gain of information. If the intervention of a role changes the procurement requirements from a vague textual description to precise technical parameters, or improves the originally missing risk clauses, the system determines that it has generated a high positive increment, which means that the collaborative behavior has high value.

[0138] Collaboration friction loss determination: The system counts invalid backtracking paths in the evolution tree in real time. When a contract is rejected back and forth between two departments without any substantial optimization, the system will identify it as serious collaboration friction. The higher the friction coefficient, the lower the communication efficiency of that link and the greater the collaboration loss.

[0139] Phase Four: Flexible Reconfiguration and Adaptive Control of Process Protocols

[0140] The system feeds back the evaluation results to the process engine in real time, enabling automatic adjustments to management methods based on their differences.

[0141] Physical acceleration of reputation path: The system builds a dynamic trust profile for each pair of collaborating roles. When the historical collaboration between the initiator and the approver has maintained high semantic increment and low friction rate, the system will determine that the link is a high-reputation path. At this time, the system automatically reconstructs the process, merges redundant approval nodes, and even opens a fast track to skip non-core links.

[0142] Dynamic reinforcement of risk points: When the system detects that the semantic increment of a link is continuously negative, or the information deviates from the initial procurement intention, it will automatically trigger adaptive control instructions. The system will dynamically add verification nodes, force supplementary supporting materials, or directly suspend the process and output a bottleneck diagnosis report to the management, indicating the root cause of the blockage or the missing information.

[0143] Phase 5. Closed-loop optimization and credit record keeping:

[0144] After the process is completed, the system saves the entire evolution trajectory and evaluation score to the collaborative credit ledger.

[0145] Continuous optimization: The evaluation data will serve as the basis for organizational structure optimization, helping companies discover which role combinations are most effective and which process nodes are poorly designed.

[0146] Immutable evidence storage: All collaboration records and system intervention instructions are stored in encrypted form. In the event of subsequent audits or quality disputes, the system can restore the true decision-making intent and collaboration process, rather than simply providing a static approval record. Specific Implementation Example 4:

[0148] like Figure 1-3 As shown, based on the content of the above specific embodiments, the following content is further disclosed, and specific use cases are provided below:

[0149] 1. Business Background:

[0150] A sudden malfunction occurred on the company's production line, requiring the urgent procurement of a high-specification vacuum pump. This task involved four roles: the production department (initiator), the technical department (specification verification), the purchasing department (business negotiation), and the finance department (budget review).

[0151] 2. Process Initiation and Initial Semantic Anchoring:

[0152] Initial state: Zhang, from the production department, submitted a request in the system, describing it as "needing one high-performance vacuum pump, the sooner the better".

[0153] System Action: The capture module captures this information and identifies it as the root node of the full-link tracing identifier. System analysis reveals that its information entropy is extremely high (both "high performance" and "the faster the better" belong to fuzzy semantics), and the initial semantic effective increment evaluation score is low.

[0154] Evolutionary tree state: The evolutionary tree generates the root node N0.

[0155] 3. Semantic increment and conflict identification in the collaboration process:

[0156] Technical intervention: After receiving the task, Li from the technical department inquired about the specific pressure parameters from the production department via instant messaging. He then added the specification to the system as "ultimate pressure ≤ 10". -3 Pa, pumping speed ≥ 50 liters / second.

[0157] System Action: The capture module identifies this semantic atom unit, the computing engine determines that the operation transforms the fuzzy description into precise parameters, the information entropy is significantly reduced, and it is determined that Li has generated an extremely high positive semantic effective increment. The evolution tree grows a child node N1.

[0158] Business Conflict: Purchasing agent Wang intervened and discovered that due to a tight delivery deadline, the original supplier was unable to deliver. He suggested changing brands and modified the supplier information in the system, but the pressure index for that brand was only 10. -2 Pa.

[0159] System Diagnosis: The system detected that Wang's modification caused a semantic conflict with Li's node N1. The evolutionary tree automatically split into parallel branches. Because this operation caused a decrease in the quality of technical parameters (semantic regression), the system recorded an increase in Wang's cooperative friction coefficient in real time.

[0160] 4. Real-time dynamic intervention and flexible process reengineering:

[0161] System decision: The system detected that the current evolution tree has serious divergence and semantic oscillation. At this time, the system query dynamic trust profile and found that Wang's collaboration friction rate in similar projects has been higher than 40% recently.

[0162] Restructuring Instruction: The system immediately triggers a risk intervention strategy. When Wang clicks "Submit", the system pops up a difference warning, requiring him to upload a "written approval letter from the technical department". At the same time, the system automatically restructures the process, dynamically injecting a "technical director countersigning" node, which forces a risk confirmation for the parameter reduction.

[0163] Positive evolution: Subsequently, Wang consulted with the technical department and finally reached a consensus at node N3: adopt the alternative solution but add an auxiliary pump, the system detected branch merging, and the semantic effective increment returned to a positive value.

[0164] 5. Trust Acceleration and Process Self-Healing:

[0165] Financial review: The process was transferred to Zhao in the finance department. The system query profile showed that Zhao and Wang in the purchasing department had maintained a high level of effective semantic increment in the past 20 collaborations and there had never been a semantic rollback (indicating that the two parties have a very high degree of tacit understanding in budget control and document compliance).

[0166] Elastic Reconfiguration: The system determines that the link is a high-reliability channel. Based on the path optimization mechanism of this invention, the system automatically executes the instruction to "skip department manager review", compressing the financial review that originally required 24 hours into a second-level automatic flow, thus achieving physical acceleration of the process.

[0167] 6. Evaluation results output and archiving:

[0168] Final output: The procurement task was completed 12 hours ahead of schedule.

[0169] System Report: After the process is completed, the system outputs a detailed dynamic evaluation report:

[0170] Li (technical): contributed the highest increment in information certainty.

[0171] Mr. Wang (procurement): A serious collaboration friction occurred, but it was corrected in time thanks to a system alert.

[0172] Process optimization suggestion: The system identified the technical parameter confirmation step as the bottleneck in this process. It is recommended that the initial form fields for "vacuum pumps" be made mandatory in the future to reduce information entropy from the source.

[0173] Evidence Preservation: The entire evolution tree and conflict resolution trajectory are encrypted and preserved as a permanent certificate of compliance for this procurement.

[0174] In conclusion, this case demonstrates that the present invention not only provides objective performance evaluation after the fact (who contributed effective information and who created friction), but also provides intelligent process control before and during the process. It changes the status quo of "people waiting for processes" and realizes "processes automatically scaling up and down according to the quality of human collaboration", ensuring that emergency procurement tasks are completed efficiently under controlled risks.

[0175] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising a reference structure" does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0176] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A dynamic evaluation system for management and procurement processes oriented towards multi-role collaboration, characterized in that, include: Sp1: The full-link interactive semantic atomization capture module is used to monitor the operation behavior of each collaborative role in the procurement process, and parse each document editing, approval opinion submission, parameter modification and instant messaging reply into semantic atomic operation units with timestamps and role identifiers; Sp2: Procurement Intent Evolution Tree Construction Module, used to construct a procurement intent evolution tree that reflects the information state changes of the procurement object in the collaboration chain, with the initial procurement demand as the root node and the semantic atomic operation unit as the growth node. The evolution tree records the complete topological trajectory of procurement information from fuzzy to precise. Sp3: Semantic Increment and Collaborative Friction Calculation Engine, used to traverse the procurement intent evolution tree, calculate the effective semantic increment value of each role node by comparing the information differences between parent and child nodes, and identify collaborative friction nodes that cause semantic regression or generate cyclic branches. SP4: Dynamic credit weighting and flexible process reconfiguration module, used to generate a collaborative performance profile of each role in real time based on historical semantic increment values ​​and the frequency of collaborative friction nodes, and dynamically adjust the current procurement process flow protocol according to the profile. The adjustment of the flow protocol includes automatic merging of approval nodes, skipping or adding mandatory verification steps. The calculation logic for the semantic effective increment value in Sp3 is as follows: the system identifies modifications made by the subsequent node to the information of the previous node as positive increments, invalid redundancy, or negative interference; when the modification of the subsequent node makes the procurement specifications more explicit and the compliance clauses more complete, it is determined to be a positive increment; when the modification of the subsequent node is revoked or repeatedly modified in subsequent steps, it is determined to be invalid redundancy; the system accumulates positive increments and deducts invalid redundancy to obtain the actual contribution value of the role in the current process, which serves as the core basis for dynamic evaluation.

2. The dynamic evaluation system for management and procurement processes oriented towards multi-role collaboration as described in claim 1, characterized in that: The process elastic reconfiguration module in Sp4 has a dynamic approval threshold drift function: the system maintains a dynamic trust library based on the collaboration efficiency profile; when it detects that the semantic increment of the historical collaboration between the initiator and the approver is continuously positive and the friction is extremely low, the system automatically increases the exemption threshold of the link to achieve automatic acceleration of low-risk processes. When the system detects that the input information of a specific role causes subsequent nodes to frequently trigger correction operations, it automatically reduces the operation privileges of that role and forcibly inserts a preprocessing verification plugin at the submission stage.

3. The dynamic evaluation system for management and procurement processes oriented towards multi-role collaboration as described in claim 1, characterized in that: The procurement intent evolution tree construction module in Sp2 includes branch pruning logic: during the process, when the system identifies that the discussion or modification of a branch is not ultimately merged into the main procurement intent, the branch is marked as a trial branch. The system tracks the participation of each role in the trial-and-error branch and uses participation as an indicator to evaluate the role's risk assessment ability, rather than a simple efficiency indicator, in order to distinguish between ineffective work and necessary risk exploration.

4. The dynamic evaluation system for management and procurement processes oriented towards multi-role collaboration as described in claim 1, characterized in that: The system also includes an implicit dependency mining module: used to analyze the high-frequency semantic atomic operation associations of different roles under non-process-defined paths; When it is found that role A's decision-making frequently depends on the informal input of role B, the system automatically establishes implicit coupling weights between A and B in the evaluation model; when evaluating A's performance, the system combines the response quality of B for weighted calculation, thereby quantifying the impact of informal collaboration on formal processes.

5. The dynamic evaluation system for management and procurement processes oriented towards multi-role collaboration as described in claim 1, characterized in that: The semantic atomization capture module in Sp1 has a cross-modal alignment function: mapping unstructured communication text to structured ERP document fields; when a commitment to price or delivery date appears in the communication text but is not updated in the ERP document, the system generates a consistency warning signal and uses this signal as a negative evaluation parameter for the degree of collaborative integrity.

6. The method for dynamic evaluation of the management procurement process executed by the system according to any one of claims 1 to 5, characterized in that, Includes the following steps: Sp101: Initialize the procurement intent evolution tree and parse the procurement requisition form into a root information node; Sp102: Capture the derived operations of each role on the root information node in real time, generate child nodes and attach them to the evolution tree; Sp103: Compare the differences between each parent-child node pair, calculate the semantic increment of the current operation, and determine whether the operation promotes process convergence or causes divergence; Sp104: Update the collaborative performance profile of the role based on the accumulated semantic increment; Sp105: Based on the updated profile, adjust the triggering conditions or approval paths of subsequent process nodes in real time and output an evaluation report.

Images