Business portfolio value calculation application method, apparatus, and medium

Abstract

This application provides a method, apparatus, and medium for calculating the value of a business portfolio, relating to the field of information technology. The method includes: obtaining a first user characteristic factor, candidate product factors, and discount strategy factors; inputting the first user characteristic factor, candidate product factors, and discount strategy factors into a multi-factor combination optimization algorithm, which, based on a multi-level cache and / or a real-time computing engine, obtains an effective business portfolio suitable for the first user and its predicted value; and recommending, accepting, and monitoring the effective business portfolio based on the predicted value. The first user characteristic factor is obtained based on the characteristics of the first user, and the effective business portfolio is a combination of at least one candidate product and / or a combination of at least one candidate product and at least one discount strategy. This application helps improve the digital and intelligent operation level of enterprises.

Description

Technical Field

[0001] This application relates at least to the field of information technology, and in particular to a method, apparatus and medium for calculating the value of a business portfolio. Background Technology

[0002] Real-time calculation of business value faces a major challenge: the trade-off between computational complexity and real-time performance. Traditional solutions typically employ a method of traversing all possible combinations and then filtering them when processing the value calculation of package combinations. For a package set containing n elements, the number of combinations grows exponentially by O(2^n) or even factorially by O(n!), which cannot complete such large-scale calculations within the millisecond-level response time required by the business, and thus cannot meet the real-time requirements of scenarios such as front-line marketing and real-time recommendations. Summary of the Invention

[0003] To address the aforementioned shortcomings, this application provides a business portfolio value calculation application method, apparatus, and medium to solve the following technical problem: how to resolve the contradiction between the complexity and real-time performance of business portfolio value calculation.

[0004] Firstly, this application provides a method for calculating the value of a business portfolio, the method comprising:

[0005] Obtain the first user characteristic factor, candidate product factor, and discount strategy factor;

[0006] The first user characteristic factor, candidate product factor and preferential strategy factor are input into the multi-factor combination optimization algorithm. The multi-factor combination optimization algorithm obtains the effective business combination suitable for the first user and its predictive value based on multi-level caching and / or real-time computing engine.

[0007] Based on predictive value, we can recommend, process, and monitor effective business portfolios.

[0008] The first user characteristic factor is obtained based on the characteristics of the first user, and the effective business combination is a combination of at least one candidate product and / or a combination of at least one candidate product with at least one preferential strategy.

[0009] Furthermore, the first user characteristic factor, candidate product factor, and discount strategy factor are obtained, specifically including:

[0010] Receive the original query request, determine whether the original query request violates the hard business rules, and obtain the valid query request that does not violate the hard business rules;

[0011] Based on a valid query request, obtain the first user's characteristic factors, including the first user's identity information, the first user's current plan, and the first user's network resources;

[0012] Based on valid query requests, candidate product factors and preferential strategy factors are obtained, including several candidate packages and several preferential activities and their business rules that meet the valid query requests.

[0013] Furthermore, the first user characteristic factors, candidate product factors, and preferential strategy factors are input into a multi-factor combination optimization algorithm. This algorithm, based on a multi-level cache and / or real-time computing engine, obtains an effective business combination suitable for the first user and its predictive value, specifically including:

[0014] The multi-factor combination optimization algorithm sorts the input candidate product factors and discount strategy factors according to the selection popularity, and obtains the first effective business combination suitable for the first user and its first predictive value from the L1 hot spot cache according to the ranking.

[0015] Among them, the first effective business combination is a first hot combination of candidate products with first candidate product factors and discount strategies with first discount strategy factors, and the second user characteristic factors of historical users who selected the first hot combination are more similar to the first user characteristic factors than the first preset threshold.

[0016] Among them, the first candidate product factor and the first discount strategy factor are the top candidate product factors and discount strategy factors ranked according to popularity.

[0017] Furthermore, the first user characteristic factors, candidate product factors, and preferential strategy factors are input into a multi-factor combination optimization algorithm. This algorithm, based on a multi-level cache and / or real-time computing engine, obtains an effective business combination suitable for the first user and its predicted value. Specifically, it also includes:

[0018] If the multi-factor combination optimization algorithm fails to obtain a first effective business combination and its first predicted value suitable for the first user from the L1 hot spot cache, it will obtain a second effective business combination and its second predicted value suitable for the first user from the L2 pre-calculated result cache according to the ranking.

[0019] Among them, the second effective business combination is a second hot combination of candidate products with second candidate product factors and discount strategies with second discount strategy factors, and the third user characteristic factor of historical users who selected the second hot combination is more similar to the first user characteristic factor than the second preset threshold.

[0020] Among them, the second candidate product factor and the second preferential strategy factor are several candidate product factors and preferential strategy factors ranked according to popularity after the first candidate product factor and the first preferential strategy factor.

[0021] Furthermore, the first user characteristic factors, candidate product factors, and preferential strategy factors are input into a multi-factor combination optimization algorithm. This algorithm, based on a multi-level cache and / or real-time computing engine, obtains an effective business combination suitable for the first user and its predicted value. Specifically, it also includes:

[0022] If the multi-factor combination optimization algorithm fails to obtain a second effective business combination and its second predicted value suitable for the first user from the L2 pre-calculation result cache, it will call the real-time calculation engine based on the first user characteristic factors, candidate product factors and preferential strategy factors to obtain a third effective business combination and its third predicted value suitable for the first user.

[0023] Among them, the third effective business combination is the third combination of candidate products with the third candidate product factor and preferential strategies with the third preferential strategy factor, and the first user characteristic factor meets the business rule constraints of the third combination.

[0024] Furthermore, based on the first user characteristic factors, candidate product factors, and preferential strategy factors, the real-time calculation engine is invoked to obtain the third effective business combination suitable for the first user and its third predicted value, specifically including:

[0025] The business rules that call the real-time computing engine to obtain candidate product factors and discount strategy factors constitute the first business rule constraint.

[0026] The real-time computing engine is invoked to obtain the second business rule constraint that the first user feature factor conforms to from the first business rule constraint;

[0027] The real-time computing engine is invoked to obtain a third combination of a third candidate product factor and a third discount strategy factor from the candidate product factors and discount strategy factors according to the second business rules;

[0028] The real-time computing engine is invoked to obtain the third predicted value of the third combination based on the historical user consumption data of the selected third combination. If the third predicted value already exists in the L3 intermediate result cache, it is retrieved directly; otherwise, the third predicted value is stored in the L3 intermediate result cache.

[0029] Furthermore, based on predicted value, the recommendation, acceptance, and monitoring of effective business portfolios are achieved, specifically including:

[0030] Based on the predicted value, including the monthly cost of the first user after selecting an effective business combination, an effective business combination is recommended to the first user based on the monthly cost.

[0031] Based on the predicted value, including the monthly cost change of the first user after selecting an effective service combination, the monthly cost change will be displayed when processing the effective service combination for the first user;

[0032] The predicted value includes the change in value of the first user after selecting an effective business combination, and the value of the work of the operations staff who handle the effective business combination for the first user is evaluated based on the change in value.

[0033] Secondly, this application provides a business portfolio value calculation application device, the device comprising:

[0034] The acquisition module is used to acquire the first user characteristic factor, candidate product factor, and discount strategy factor;

[0035] The calculation module, connected to the acquisition module, is used to input the first user characteristic factors, candidate product factors and preferential strategy factors into the multi-factor combination optimization algorithm. The multi-factor combination optimization algorithm obtains the effective business combination suitable for the first user and its predictive value based on multi-level caching and / or real-time computing engine.

[0036] The application module, connected to the computing module, is used to recommend, accept, and monitor effective business combinations based on predicted value.

[0037] The first user characteristic factor is obtained based on the characteristics of the first user, and the effective business combination is a combination of at least one candidate product and / or a combination of at least one candidate product with at least one preferential strategy.

[0038] Thirdly, this application provides a computer device including a processor and a memory, wherein the memory stores a computer program, and when the processor runs the computer program stored in the memory, the processor executes the business portfolio value calculation application method as described above.

[0039] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the business portfolio value calculation application method as described above.

[0040] This application provides a business portfolio value calculation application method, device, and medium. By acquiring multiple factors for portfolio optimization, it selects effective business portfolios suitable for users and predicts their value, thereby realizing business recommendation, acceptance, and monitoring, which helps improve the digital and intelligent operation level of enterprises. Attached Figure Description

[0041] Figure 1 This is a flowchart of a business portfolio value calculation application method according to an embodiment of this application;

[0042] Figure 2 This is an architecture diagram of a business portfolio value calculation application system according to an embodiment of this application;

[0043] Figure 3This is a schematic diagram illustrating the composition of a multi-factor embodiment of this application;

[0044] Figure 4 This is a schematic diagram of a hard rule filtering method according to an embodiment of this application;

[0045] Figure 5 This is a flowchart of a multi-factor combinatorial optimization algorithm according to an embodiment of this application;

[0046] Figure 6 This is a schematic diagram of a rule-based pruning implementation according to an embodiment of this application;

[0047] Figure 7 This is an interaction sequence diagram of a business portfolio value calculation application method according to an embodiment of this application;

[0048] Figure 8 This is a flowchart of a real-time computing engine training method according to an embodiment of this application;

[0049] Figure 9 This is a flowchart of another business portfolio value calculation application method according to an embodiment of this application;

[0050] Figure 10 This is a schematic diagram of the structure of a business portfolio value calculation application device according to an embodiment of this application;

[0051] Figure 11 This is a schematic diagram of the structure of a computer device according to an embodiment of this application;

[0052] Figure 12 This is a schematic diagram of the structure of a computer-readable storage medium according to an embodiment of this application. Detailed Implementation

[0053] To enable those skilled in the art to better understand the technical solution of this application, the embodiments of this application will be further described in detail below with reference to the accompanying drawings.

[0054] It is understood that the specific embodiments and accompanying drawings described herein are merely for explaining this application and are not intended to limit this application.

[0055] It is understood that, without conflict, the various embodiments and features in the embodiments of this application can be combined with each other.

[0056] It is understood that, for ease of description, only the parts relevant to this application are shown in the accompanying drawings, while parts unrelated to this application are not shown in the drawings.

[0057] It is understood that each module or unit involved in the embodiments of this application may correspond to only one entity structure, or may be composed of multiple entity structures, or multiple modules or units may be integrated into one entity structure.

[0058] It is understood that, without conflict, the functions and steps indicated in the flowcharts and block diagrams of this application may occur in a different order than that indicated in the accompanying drawings.

[0059] It is understood that the flowcharts and block diagrams of this application illustrate the possible architecture, functions, and operations of systems, apparatuses, devices, and methods according to various embodiments of this application. Each block in a flowchart or block diagram may represent a module, unit, program segment, or code, containing executable instructions for implementing the specified function. Furthermore, each block or combination of blocks in the block diagrams and flowcharts may be implemented using a hardware-based device to implement the specified function, or using a combination of hardware and computer instructions.

[0060] It is understood that the modules and units involved in the embodiments of this application can be implemented by software or by hardware. For example, the modules and units can be located in the processor.

[0061] Example 1:

[0062] like Figure 1 As shown, this application provides a method for calculating the value of a business portfolio, the method comprising:

[0063] S1. Obtain the first user characteristic factor, candidate product factor, and discount strategy factor;

[0064] S2. Input the first user characteristic factor, candidate product factor and discount strategy factor into the multi-factor combination optimization algorithm. The multi-factor combination optimization algorithm obtains the effective business combination suitable for the first user and its predictive value based on multi-level caching and / or real-time computing engine.

[0065] S3. Recommend, accept, and monitor effective business combinations based on predicted value;

[0066] The first user characteristic factor is obtained based on the characteristics of the first user, and the effective business combination is a combination of at least one candidate product and / or a combination of at least one candidate product with at least one preferential strategy.

[0067] In this embodiment, the provided method optimizes the combination of multiple factors, filters effective business combinations suitable for users and predicts their value, thereby enabling business recommendation, acceptance and monitoring, which helps improve the digital and intelligent operation level of enterprises.

[0068] Specifically, this embodiment provides a nested (package stacking) combination optimization algorithm and caching collaboration system for real-time computing across all scenarios. This embodiment belongs to the field of enterprise digital operation technology, with the core objective of improving operational efficiency and decision-making accuracy through data-driven approaches. Specifically, it involves a real-time value calculation technology for high-concurrency, multi-constraint business scenarios. Especially in industries such as telecommunications and finance, enterprise operating systems need to process massive amounts of user order data in real time (daily increment ≥ PB level), dynamically integrating multi-dimensional factors such as product business, preferential strategies, and user behavior to support T+1 value calculation and T-1 retrospective analysis, ultimately achieving dynamic inclusion of business rules, real-time calculation of business value, and real-time benchmarking of business resources.

[0069] More specifically, this embodiment describes a system and method for processing massive package combinations and real-time calculation of order value. Real-time calculation of such business value mainly faces the following bottlenecks and shortcomings:

[0070] (1) The contradiction between computational complexity and real-time performance: When processing the value calculation of package combinations, traditional solutions usually adopt the method of traversing all possible combinations and then filtering. For a package set containing n elements, the number of combinations grows exponentially by O(2^n) or even factorial O(n!), which cannot complete such a large-scale calculation within the millisecond response time required by the business, and cannot meet the real-time requirements of scenarios such as front-line marketing and real-time recommendation.

[0071] (2) Constraints of business rule complexity: Existing calculation models are difficult to efficiently integrate complex and dynamically changing business rules (such as mutually exclusive packages, user level privileges, etc.). They usually adopt post-filtering (calculate first and then remove) or simple static rule judgment, which generates a lot of invalid calculations.

[0072] (3) Rigid caching strategy: Most systems adopt general caching strategies such as LRU (least recently used, a cache eviction algorithm) without optimizing for the hot features of business queries, resulting in high-value calculation results not being able to reside in the cache for a long time and a low cache hit rate (usually less than 50%).

[0073] This embodiment aims to overcome the challenges of collaborative optimization in real-time penetration of dynamic rules, real-time computation of ultra-large-scale combinations, and adaptive scheduling of caching strategies. Ultimately, it achieves real-time computation of business value, supports T+1 prediction and churn warning, takes data standardization governance as its foundation, AI real-time computation as its core, and business decision-making closed loop as its exit, and solves two major pain points: "uncontrollable impact of business changes" and "difficulty in quantifying package value".

[0074] The core of this embodiment lies in constructing a three-tiered pipelined processing architecture. This architecture decomposes complex computational tasks into three collaborative layers: serialization modeling, a pre-computation engine, and a caching strategy. This achieves ultimate optimization from request to response. The system architecture is as follows: Figure 2 As shown.

[0075] This embodiment presents a nested combination optimization method for real-time computing across all scenarios, comprising: receiving user calculation requests and extracting user features, package factors, and discount strategies; performing compliance verification on request factors based on a pre-set business constraint matrix, directly intercepting invalid requests; for valid requests, executing a multi-factor nested combination algorithm: traversing the product digitization model and performing real-time verification at each traversal node according to the business constraint matrix, pruning invalid paths in real time, and only performing value calculation on valid paths; returning the calculation results and caching the results according to a predefined strategy. Before the "execute multi-factor nested combination algorithm" step, the calculation factors are prioritized according to the strength of their business constraints, with the factors with the strongest constraints being processed first. A multi-level caching collaboration strategy is adopted, storing highly popular calculation results in the L1 hot spot cache; storing results generated by offline pre-computation jobs in the L2 pre-computation cache; and storing common intermediate results generated during the calculation process in the L3 common factor cache. A machine learning model is used to analyze query logs and generate a popularity weight model; based on the popularity weight model, the priority and eviction strategy of data in each level of cache are dynamically adjusted.

[0076] In one embodiment, S1, obtaining the first user characteristic factor, candidate product factor, and discount strategy factor specifically includes:

[0077] Receive the original query request, determine whether the original query request violates the hard business rules, and obtain the valid query request that does not violate the hard business rules;

[0078] Based on a valid query request, obtain the first user's characteristic factors, including the first user's identity information, the first user's current plan, and the first user's network resources;

[0079] Based on valid query requests, candidate product factors and preferential strategy factors are obtained, including several candidate packages and several preferential activities and their business rules that meet the valid query requests.

[0080] In this embodiment, as Figure 2 As shown, the first phase of this system is the serialization model layer, which implements request filtering and standardization.

[0081] This layer is the system's entry gateway, responsible for receiving raw query requests and performing initial invalid request interception and standardization, as well as implementing business rule constraint matrix verification. Specifically, it includes:

[0082] (1) The system has a built-in multidimensional constraint matrix based on business rules (such as package mutual exclusion and user level qualification), which predefines the compatibility relationship between different dimension factors.

[0083] (2) When a query request arrives, first extract the key factors in the request (such as target package: package A) and quickly compare them with the constraint matrix.

[0084] (3) Determine whether the request violates hard business rules within milliseconds (e.g., the request contains mutually exclusive packages A and B at the same time). For invalid requests, directly intercept and return, avoiding unnecessary consumption of computing resources caused by penetration into the backend.

[0085] like Figure 3 As shown, the multi-factor includes user tags, product tags, income tags, resource tags, etc. For user characteristic factors, it can include user identity information, consumption level, current package type, current monthly fee, broadband / terminal / service information, etc. For product factors, it can include package type, package monthly fee, package required resources, package available discounts, etc. For discount strategy factors, it can include discount strength and stacking method, etc. The resource value is calculated by combining these factors.

[0086] like Figure 4 As shown, invalid request judgment is achieved by filtering key factors (such as 0, 1, 2, 3) from the request and quickly comparing them with the constraint matrix to determine whether there is a violation of hard business rules, such as 0 and 3 being mutually exclusive packages A and B.

[0087] In one embodiment, S2, the first user characteristic factor, candidate product factor, and discount strategy factor are input into a multi-factor combination optimization algorithm. The multi-factor combination optimization algorithm obtains an effective business combination suitable for the first user and its predicted value based on a multi-level cache and / or real-time computing engine, specifically including:

[0088] The multi-factor combination optimization algorithm sorts the input candidate product factors and discount strategy factors according to the selection popularity, and obtains the first effective business combination suitable for the first user and its first predictive value from the L1 hot spot cache according to the ranking.

[0089] Among them, the first effective business combination is a first hot combination of candidate products with first candidate product factors and discount strategies with first discount strategy factors, and the second user characteristic factors of historical users who selected the first hot combination are more similar to the first user characteristic factors than the first preset threshold.

[0090] Among them, the first candidate product factor and the first discount strategy factor are the top candidate product factors and discount strategy factors ranked according to popularity.

[0091] In this embodiment, as Figure 2 As shown, the second stage of this system is the pre-computation engine layer, which performs core algorithm processing.

[0092] For valid requests that pass the verification, they enter the core computing layer and are processed by the multi-factor nested combination algorithm to implement the multi-factor nested combination algorithm.

[0093] The multi-factor nested combination algorithm is the core technology, and its detailed internal processing flow is as follows: Figure 5 As shown, it includes:

[0094] Input: Serialized user characteristics, vectorized factors in the current product digitization model, and a set of discount strategies.

[0095] Process: The algorithm does not generate all possible permutations, but rather performs intelligent traversal based on dynamic programming and constraint propagation strategies, such as:

[0096] (1) Factor priority sorting: Based on historical data or business weight, the calculation factors are sorted and the factors with the strongest constraints are processed first to quickly narrow down the search space.

[0097] (2) Deep association mapping: Traverse the product digital model and dynamically construct a combination tree by utilizing the relationship between factors.

[0098] (3) Real-time invalid combination removal: At each step of the combination construction, the constraint matrix is ​​used for verification. Once a path is found to violate the rules, it is pruned immediately, and all subsequent calculations under that path are terminated. The pruning effect is as follows: Figure 6 As shown.

[0099] The core class of the pre-computation engine is responsible for executing the multi-factor nested combination algorithm, performing real-time calculations and pruning, and is implemented as follows:

[0100] class PrecomputationEngine:

[0101] def __init__(self, product_model, constraint_matrix):

[0102] self.product_model = product_model # Digital product model

[0103] self.constraint_matrix = constraint_matrix # Business constraint matrix

[0104] self.intermediate_cache = {} # Intermediate result cache (L3)

[0105] def find_optimal_combinations(self, user_features):

[0106] # 1. Obtain relevant factors

[0107] factors = self.product_model.get_factors(user_features)

[0108] optimal_plans = []

[0109] # 2. Dynamic Programming-Based Traversal and Combinatorial Approach

[0110] for factor in factors:

[0111] # Generate candidate combinations (actually, this should be a state transition)

[0112] candidate = self._generate_combination(factor)

[0113] # 3. Key Step: Real-time Pruning

[0114] if not self.constraint_matrix.is_valid(candidate):

[0115] continue # Invalid combinations are discarded immediately

[0116] # 4. Calculate portfolio value

[0117] value = self._calculate_value(candidate)

[0118] optimal_plans.append({'factors': candidate, 'value':value})

[0119] # 5. Return the optimal result

[0120] return sorted(optimal_plans, key=lambda x: x['value'],reverse=True)[:5]

[0121] def _calculate_value(self, combination):

[0122] # Calculate combined value - Support intermediate result caching: Check intermediate cache

[0123] cache_key = hash(tuple(combination))

[0124] if cache_key in self.intermediate_cache:

[0125] return self.intermediate_cache[cache_key]

[0126] # Simulate complex calculations (monthly rent, discounts, costs, etc.)

[0127] value = sum(self.product_model.get_factor_value(factor) forfactor in combination)

[0128] # Store intermediate results

[0129] self.intermediate_cache[cache_key] = value

[0130] return value

[0131] By leveraging dynamic programming and pruning to efficiently search for the optimal solution and reusing intermediate computational results, the computational complexity is reduced from the traditional O(n!) to approximately O(n!). 2 This represents a fundamental leap from "uncomputable" to "computable in real time".

[0132] like Figure 2 As shown, Phase 3 of this system: the multi-level cache collaboration layer achieves efficient result return.

[0133] To ensure low latency under high concurrency, a three-layer caching coordination mechanism was designed:

[0134] (1) L1 - Hotspot cache (Redis): Stores the final calculation results of the most frequently queried queries, with a very short response time (<5ms) and adopts an elimination strategy based on hot weight.

[0135] (2) L2 - Pre-calculation result cache (Redis): Stores the results generated by offline pre-calculation jobs. The system calculates the possible results of popular package combinations in advance during off-peak business periods (such as at night) and stores them in this layer.

[0136] (3) L3 - Common Factor Cache: Stores intermediate calculation results (such as the calculation result of a specific coupon). Multiple final results may share the same intermediate factor. Cache such results to avoid duplicate calculations.

[0137] (4) Cache query strategy: For a request, the system queries in the order of "L1 -> L2 -> Calculation engine -> Backfill cache". If L1 and L2 are not hit, but some intermediate results can be found in L3, the calculation engine only needs to calculate the remaining part, which greatly improves efficiency.

[0138] In one embodiment, S2, the first user characteristic factor, candidate product factor, and discount strategy factor are input into a multi-factor combination optimization algorithm. The multi-factor combination optimization algorithm obtains an effective business combination suitable for the first user and its predicted value based on a multi-level cache and / or real-time computing engine. Specifically, it also includes:

[0139] If the multi-factor combination optimization algorithm fails to obtain a first effective business combination and its first predicted value suitable for the first user from the L1 hot spot cache, it will obtain a second effective business combination and its second predicted value suitable for the first user from the L2 pre-calculated result cache according to the ranking.

[0140] Among them, the second effective business combination is a second hot combination of candidate products with second candidate product factors and discount strategies with second discount strategy factors, and the third user characteristic factor of historical users who selected the second hot combination is more similar to the first user characteristic factor than the second preset threshold.

[0141] Among them, the second candidate product factor and the second preferential strategy factor are several candidate product factors and preferential strategy factors ranked according to popularity after the first candidate product factor and the first preferential strategy factor.

[0142] In this embodiment, the multi-level cache coordination layer is responsible for managing the query strategies and data synchronization of the three-level cache (L1 / L2 / L3), and is implemented as follows:

[0143] class MultiLevelCache:

[0144] def __init__(self):

[0145] # Initialize the three-level cache client

[0146] self.l1_cache = {} # Hotspot caching (simulated by memory dictionary)

[0147] self.l2_cache = {} # Cache pre-computed results

[0148] self.l3_cache = {} # Public factor cache

[0149] self.access_count = {} # Access popularity statistics

[0150] def get(self, key):

[0151] # Cache Query: 1. Query L1 hotspot cache

[0152] If the key is in self.l1_cache:

[0153] self._update_heat(key)

[0154] return self.l1_cache[key]

[0155] # 2. Query L2 pre-computed cache

[0156] If the key is in self.l2_cache:

[0157] value = self.l2_cache[key]

[0158] self._async_backfill_l1(key, value) # Asynchronous backfill L1

[0159] return value

[0160] # 3. Query L3 intermediate result cache

[0161] If the key is in self.l3_cache:

[0162] return self.l3_cache[key]

[0163] # All caches were missed

[0164] return None

[0165] def set(self, key, value, level='L2'):

[0166] # Cache write - Specify storage level

[0167] if level == 'L1':

[0168] self.l1_cache[key] = value

[0169] elif level == 'L2':

[0170] self.l2_cache[key] = value

[0171] elif level == 'L3':

[0172] self.l3_cache[key] = value

[0173] def _update_heat(self, key):

[0174] # Update access popularity - Used to dynamically adjust caching strategies

[0175] self.access_count[key] = self.access_count.get(key, 0) + 1

[0176] # Dynamically increase cache level based on popularity

[0177] if self.access_count[key] > 10: # Popularity threshold

[0178] if key in self.l2_cache and key not in self.l1_cache:

[0179] self.l1_cache[key] = self.l2_cache[key]

[0180] def _async_backfill_l1(self, key, value):

[0181] # Asynchronous L1 cache backfilling - Production environments should be truly asynchronous.

[0182] self.l1_cache[key] = value

[0183] In one embodiment, S2, the first user characteristic factor, candidate product factor, and discount strategy factor are input into a multi-factor combination optimization algorithm. The multi-factor combination optimization algorithm obtains an effective business combination suitable for the first user and its predicted value based on a multi-level cache and / or real-time computing engine. Specifically, it also includes:

[0184] If the multi-factor combination optimization algorithm fails to obtain a second effective business combination and its second predicted value suitable for the first user from the L2 pre-calculation result cache, it will call the real-time calculation engine based on the first user characteristic factors, candidate product factors and preferential strategy factors to obtain a third effective business combination and its third predicted value suitable for the first user.

[0185] Among them, the third effective business combination is the third combination of candidate products with the third candidate product factor and preferential strategies with the third preferential strategy factor, and the first user characteristic factor meets the business rule constraints of the third combination.

[0186] In this embodiment, Figure 7 To achieve efficient computation of the interactive sequence graph using multi-level caching and real-time computing engines, if the computation result can be obtained directly from the cache, there is no need for real-time computation. In the real-time computation process, intermediate factors can also be used to speed up the computation, thereby improving computational efficiency and saving computational resources.

[0187] In one embodiment, a real-time calculation engine is invoked based on the first user characteristic factors, candidate product factors, and discount strategy factors to obtain a third effective business combination suitable for the first user and its third predicted value, specifically including:

[0188] The business rules that call the real-time computing engine to obtain candidate product factors and discount strategy factors constitute the first business rule constraint.

[0189] The real-time computing engine is invoked to obtain the second business rule constraint that the first user feature factor conforms to from the first business rule constraint;

[0190] The real-time computing engine is invoked to obtain a third combination of a third candidate product factor and a third discount strategy factor from the candidate product factors and discount strategy factors according to the second business rules;

[0191] The real-time computing engine is invoked to obtain the third predicted value of the third combination based on the historical user consumption data of the selected third combination. If the third predicted value already exists in the L3 intermediate result cache, it is retrieved directly; otherwise, the third predicted value is stored in the L3 intermediate result cache.

[0192] In this embodiment, as Figure 2 As shown, the fourth stage of this system is the background optimization layer, which enables system self-optimization.

[0193] Offline pre-computation module: Periodically runs batch jobs to simulate a massive number of possible combinations, and pre-calculates high-frequency results and stores them in the L2 cache, as follows:

[0194] # The pre-computation engine layer and the multi-level cache coordination layer work together: initialization

[0195] product_model = ProductDigitalModel() # Product digital model

[0196] constraint_matrix = ConstraintMatrix() # Business constraint matrix

[0197] cache_coordinator = MultiLevelCacheCoordinator() # Cache coordinator

[0198] precomputation_engine = PrecomputationEngine(product_model,constraint_matrix) # Precomputation engine

[0199] def get_optimal_plan(user_request: Dict) -> List[Dict]:

[0200] # Business Interface: Requests the optimal package combination for users.

[0201] # 1. Generate a unique cache key based on user characteristics and request parameters.

[0202] cache_key = generate_cache_key(user_request)

[0203] # 2. Query multi-level cache

[0204] cached_result = cache_coordinator.get(cache_key)

[0205] if cached_result is not None:

[0206] return cached_result # If the cache is hit, return directly.

[0207] # 3. Cache miss, call the pre-computation engine for real-time calculation.

[0208] optimal_plans = precomputation_engine.find_optimal_combinations(user_request)

[0209] # 4. Asynchronously write the calculation results to the cache (determine whether to store in L2 or L1 based on the strategy): If the result is calculated for a common user group, store it in L2.

[0210] cache_coordinator.set(cache_key, optimal_plans, cache_level='L2')

[0211] #5. Simultaneously, some intermediate calculation results (such as the value of a single factor) can be asynchronously stored in L3.

[0212] _async_store_intermediate_results(precomputation_engine.intermediate_results_cache)

[0213] return optimal_plans

[0214] Adaptive learning module: such as Figure 8 As shown, the machine learning model continuously analyzes query logs, dynamically identifies changes in business hotspots, and adjusts the hotspot weight model to guide the eviction strategy of L1 cache and the priority of pre-computed jobs, enabling the system to have self-optimization capabilities.

[0215] Dynamic threshold adjustment: Risk threshold is dynamically calculated based on historical churn data: Risk threshold = f(user value, competitive landscape, seasonal factors, etc.)`.

[0216] In one implementation, S3, recommending, accepting, and monitoring effective business combinations based on predicted value, specifically includes:

[0217] Based on the predicted value, including the monthly cost of the first user after selecting an effective business combination, an effective business combination is recommended to the first user based on the monthly cost.

[0218] Based on the predicted value, including the monthly cost change of the first user after selecting an effective service combination, the monthly cost change will be displayed when processing the effective service combination for the first user;

[0219] The predicted value includes the change in value of the first user after selecting an effective business combination, and the value of the work of the operations staff who handle the effective business combination for the first user is evaluated based on the change in value.

[0220] In this embodiment, as Figure 9 As shown, to illustrate the practical application of this embodiment, the following example is provided:

[0221] Scenario: The user is considering changing their mobile phone plan.

[0222] Basic information: Currently a 4G plan user with a monthly fee of 58 yuan.

[0223] Objective: Smart home delivery service, allowing users to view and switch to a more cost-effective 5G plan through the app.

[0224] In this practical application example:

[0225] Phase 1: Data Standardization Governance

[0226] (1) The system standardizes the data on packages, tariffs, and promotional activities across the entire network and assigns them unified tags. For example:

[0227] ① Xiao Wang's current plan: [4G migration, consumption level: low to medium, plan ID: P4G001, monthly fee: 58, data: 20GB...]

[0228] ② Candidate 5G Plan: [5G Entry-level, Plan ID: P5G001, Monthly Fee: 79, Data: 30GB...]

[0229] ③Promotion: [Discount ID: Pm001, Promotion restrictions: First year discount, bundled package, Promotion rules: 20 RMB off the first year's monthly fee]

[0230] (2) All of these data are constructed into a unified “product digital model” and “business constraint matrix” (e.g., rules specify which activities can be combined with P5G001 and which packages cannot be shared with discounts).

[0231] Phase Two: Intelligent Computing Layer

[0232] When a user selects "Upgrade Package" or "Recommended Package" on the app, the system's engine starts working:

[0233] (1) Serialization model layer (request filtering):

[0234] Upon receiving a user's request, the system immediately verifies it using a business constraint matrix. If the user is a "4G migration" user, the system automatically filters out all applicable "4G to 5G upgrade" packages and offers, and invalid options (such as packages limited to new users) are directly excluded.

[0235] (2) Pre-computation engine layer (precision calculation and pruning):

[0236] The system activates the "multi-factor nested combination algorithm" to generate possible meal combinations for Xiao Wang. For example:

[0237] Combination A: P5G001 + Pm001 (First year discount)

[0238] Combination B: P5G002 (No discount)

[0239] Combo C:P5G003 + Bonus Offer (Includes Video Membership)

[0240] During the generation process, the algorithm prunes in real time to remove those that conflict with the user's current plan.

[0241] Calculate the precise T+1 predicted value for each valid combination: that is, the monthly fee for Xiao Wang after he selects a package.

[0242] (3) Multi-level cache coordination layer (ultra-fast response):

[0243] Before performing the calculation, the system first queries the cache. Since "4G migration users" is a common group, the pre-calculated results for the high-frequency combination P5G001 + Pm001 are likely already stored in the L2 cache.

[0244] The system instantly retrieves the result from the cache and returns it directly to the front end, eliminating the need for repeated calculations for common requests.

[0245] Phase Three: Implementation of the Decision-Making Closed Loop (Demonstration of Business Value)

[0246] The calculation results are applied to specific business processes:

[0247] (1) Recommendation before acceptance:

[0248] The app interface clearly displays the comparison results of various resources for the recommended combination A: 5G Enjoy Package (79 yuan) + 4G Migration Exclusive Gift (first year monthly fee only 59 yuan).

[0249] The interface also displays a chart of monthly expenses predicted for T+1, providing a clear overview of future expenditures and greatly improving conversion rates.

[0250] (2) Monthly rent reminder during processing:

[0251] Before order acceptance and pre-submission, you can also click to view the comparison of package resources and monthly fees before and after acceptance. At the same time, the system will pop up a window to remind you: "The monthly fee of the package you selected has increased by 20 yuan compared to the previous month's monthly fee. Please note." This provides an immediate notification of the acceptance status and avoids future customer complaints.

[0252] (3) Post-acceptance report monitoring (T-1 retrospective and control):

[0253] The following day, managers can monitor overall business value changes on the operations dashboard.

[0254] T-1 Retrospective: The report shows which packages were recommended and subscribed to yesterday through this function. Based on the T-1 package, it is estimated that the monthly revenue will increase by ** yuan, and users who change packages will be strictly controlled.

[0255] Monitoring and Alerts: The dashboard also monitors "invalid value enhancement" behavior, guiding operations staff to focus on effective value enhancement.

[0256] The advantages of the technical solution in this embodiment compared with the traditional solution are shown in Table 1 below:

[0257]

[0258] This embodiment implements a "multi-factor nested combination algorithm," which, based on inputs such as business constraints (mutually exclusive packages, etc.) and discount strategies, traverses the product digitization model to generate all permutations and combinations, and removes invalid combinations in real time. Combined with offline pre-computation and a multi-level caching collaboration mechanism (such as Redis layered caching), it achieves real-time calculation of group monthly fees for hundreds of millions of orders (<100ms response time). A three-level pipeline architecture of "serialization model - pre-computation engine - caching strategy" is designed to support high-concurrency queries. Ultimately, it overcomes the bottleneck of traditional combinatorial optimization algorithms that cannot simultaneously address real-time performance and complex business rules, while reducing redundant computation by 90% through pre-computation and caching.

[0259] Specifically:

[0260] (1) Real-time computational pruning mechanism based on constraint matrix: business rules (mutual exclusion, dependency, regional restriction) are abstracted into a computable three-dimensional constraint matrix, and real-time verification and path pruning are performed during the traversal of the multi-factor nested combination algorithm, which fundamentally avoids the generation and computation of invalid combinations.

[0261] (2) Dynamic collaborative strategy of offline pre-computation and multi-level caching: A collaborative architecture consisting of L1 (hot spot), L2 (pre-computation) and L3 (common factor) caches was designed and linked with the offline pre-computation job, rather than simply caching the results. The system can intelligently preheat the pre-computation results to L2, promote the high-frequency results to L1, and settle the intermediate calculation results to L3, so as to maximize the distribution of the computing load.

[0262] (3) Adaptive caching optimization method based on business popularity: By analyzing real-time query logs through machine learning models, dynamic identification and prediction of changes in business hotspots are made, and the popularity weight and eviction strategy of data in the cache are adjusted accordingly, so that the cached content always maintains a high degree of matching with the current business needs, thereby continuously maintaining a high hit rate.

[0263] (4) "Serialization-Computation-Cache" three-level pipeline processing architecture: The request processing process is standardized and modularized into three clear and collaborative layers: serialization layer (responsible for verification and interception), computation layer (responsible for core algorithm processing), and cache layer (responsible for efficient return), which realizes the system's high cohesion, low coupling and high scalability.

[0264] Example 2:

[0265] like Figure 10 As shown, this application provides a business portfolio value calculation application device, the device comprising:

[0266] Module 1 is used to acquire the first user characteristic factor, candidate product factor, and discount strategy factor;

[0267] Calculation module 2, connected to acquisition module 1, is used to input the first user feature factors, candidate product factors and preferential strategy factors into the multi-factor combination optimization algorithm. The multi-factor combination optimization algorithm obtains the effective business combination suitable for the first user and its predictive value based on multi-level caching and / or real-time computing engine.

[0268] Application module 3, connected to computing module 2, is used to recommend, accept, and monitor effective business combinations based on predicted value;

[0269] The first user characteristic factor is obtained based on the characteristics of the first user, and the effective business combination is a combination of at least one candidate product and / or a combination of at least one candidate product with at least one preferential strategy.

[0270] In one embodiment, the acquisition module 1 specifically includes:

[0271] The receiving unit is used to receive the original query request, determine whether the original query request violates the hard business rules, and obtain the valid query request that does not violate the hard business rules.

[0272] The user feature unit, connected to the receiving unit, is used to obtain the first user feature factor according to a valid query request, including the first user's identity information, the first user's current package, and the first user's network resources.

[0273] The product discount unit, connected to the receiving unit, is used to obtain candidate product factors and discount strategy factors based on valid query requests, including several candidate packages and several discount activities and their business rules that meet the valid query requests.

[0274] In one embodiment, the calculation module 2 specifically includes:

[0275] The first cache unit is used by the multi-factor combination optimization algorithm to sort the input candidate product factors and discount strategy factors according to the selection popularity, and to obtain the first effective business combination suitable for the first user and its first predicted value from the L1 hot spot cache according to the sorting.

[0276] Among them, the first effective business combination is a first hot combination of candidate products with first candidate product factors and discount strategies with first discount strategy factors, and the second user characteristic factors of historical users who selected the first hot combination are more similar to the first user characteristic factors than the first preset threshold.

[0277] Among them, the first candidate product factor and the first discount strategy factor are the top candidate product factors and discount strategy factors ranked according to popularity.

[0278] In one embodiment, the calculation module 2 further includes:

[0279] The second cache unit is used to obtain, according to the sorting, a second effective business combination and its second predicted value suitable for the first user from the L2 pre-calculation result cache if the multi-factor combination optimization algorithm does not obtain a first effective business combination and its first predicted value suitable for the first user from the L1 hot spot cache.

[0280] Among them, the second effective business combination is a second hot combination of candidate products with second candidate product factors and discount strategies with second discount strategy factors, and the third user characteristic factor of historical users who selected the second hot combination is more similar to the first user characteristic factor than the second preset threshold.

[0281] Among them, the second candidate product factor and the second preferential strategy factor are several candidate product factors and preferential strategy factors ranked according to popularity after the first candidate product factor and the first preferential strategy factor.

[0282] In one embodiment, the calculation module 2 further includes:

[0283] The real-time computing unit is used to obtain a second effective business combination and its second predicted value suitable for the first user if the multi-factor combination optimization algorithm does not obtain a second effective business combination and its second predicted value suitable for the first user from the L2 pre-calculation result cache. Then, it calls the real-time computing engine based on the first user characteristic factors, candidate product factors and preferential strategy factors to obtain a third effective business combination and its third predicted value suitable for the first user.

[0284] Among them, the third effective business combination is the third combination of candidate products with the third candidate product factor and preferential strategies with the third preferential strategy factor, and the first user characteristic factor meets the business rule constraints of the third combination.

[0285] In one embodiment, the real-time computing unit specifically includes:

[0286] The rule unit, which is used to call the real-time computing engine to obtain candidate product factors and discount strategy factors, constitutes the first business rule constraint;

[0287] The rule filtering unit, connected to the rule unit, is used to call the real-time computing engine to obtain the second business rule constraint that the first user feature factor conforms to from the first business rule constraint;

[0288] The combination unit, connected to the rule filtering unit, is used to call the real-time calculation engine to obtain the third combination of the third candidate product factor and the third discount strategy factor from the candidate product factor and the discount strategy factor according to the second business rule;

[0289] The third cache unit, connected to the combination unit, is used to call the real-time computing engine to obtain the third prediction value of the third combination based on the historical user consumption of the selected third combination. If the third prediction value already exists in the L3 intermediate result cache, it is directly called to obtain it; otherwise, the third prediction value is stored in the L3 intermediate result cache.

[0290] In one embodiment, application module 3 specifically includes:

[0291] The recommendation unit is used to recommend effective service combinations to the first user based on the predicted value, including the first user's monthly cost after selecting an effective service combination;

[0292] The processing unit is used to prompt the monthly fee change when processing the effective service combination for the first user, based on the predicted value, including the monthly fee change of the first user after selecting the effective service combination;

[0293] The monitoring unit is used to evaluate the work value of the operations personnel who handle the effective business combination for the first user based on the predicted value, including the value change of the first user after selecting an effective business combination.

[0294] Example 3:

[0295] like Figure 11 As shown, Embodiment 3 of this application provides a computer device, which includes a memory and a processor. The memory stores a computer program. When the processor runs the computer program stored in the memory, the processor executes the business portfolio value calculation application method as described in Embodiment 1. This computer device can be the business portfolio value calculation application device as described in Embodiment 2.

[0296] The memory is connected to the processor. The memory can be flash memory, read-only memory or other types of memory. The processor can be a central processing unit or a microcontroller.

[0297] Example 4:

[0298] like Figure 12 As shown, Embodiment 4 of this application provides a computer-readable storage medium storing a computer program. When the computer program is run by a processor, it implements the business portfolio value calculation application method as described in Embodiment 1.

[0299] The computer-readable storage medium includes volatile or non-volatile, removable or non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, computer program units, or other data). Computer-readable storage media include, but are not limited to, RAM (Random Access Memory), ROM (Read-Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), flash memory or other memory technologies, CD-ROM (Compact Disc Read-Only Memory), DVD or other optical disc storage, cartridges, magnetic tapes, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer.

[0300] Embodiments 1-4 of this application provide a business portfolio value calculation application method, apparatus and medium. By acquiring multiple factors for portfolio optimization, effective business portfolios suitable for users are screened and their value is predicted. Based on this, business recommendation, acceptance and monitoring are realized, which helps to improve the digital and intelligent operation level of enterprises.

[0301] It is understood that the above embodiments are merely exemplary embodiments used to illustrate the principles of this application, and this application is not limited thereto. For those skilled in the art, various modifications and improvements can be made without departing from the spirit and substance of this application, and these modifications and improvements are also considered to be within the scope of protection of this application.

Claims

1. A method for calculating the value of a business portfolio, characterized in that, The method includes: Obtain the first user characteristic factor, candidate product factor, and discount strategy factor; The first user characteristic factor, candidate product factor and preferential strategy factor are input into the multi-factor combination optimization algorithm. The multi-factor combination optimization algorithm obtains the effective business combination suitable for the first user and its predictive value based on multi-level caching and / or real-time computing engine. Based on predictive value, we can recommend, process, and monitor effective business portfolios. The first user characteristic factor is obtained based on the characteristics of the first user, and the effective business combination is a combination of at least one candidate product and / or a combination of at least one candidate product with at least one preferential strategy.

2. The method according to claim 1, characterized in that, Obtain the first user characteristic factor, candidate product factor, and discount strategy factor, specifically including: Receive the original query request, determine whether the original query request violates the hard business rules, and obtain the valid query request that does not violate the hard business rules; Based on a valid query request, obtain the first user's characteristic factors, including the first user's identity information, the first user's current plan, and the first user's network resources; Based on valid query requests, candidate product factors and preferential strategy factors are obtained, including several candidate packages and several preferential activities and their business rules that meet the valid query requests.

3. The method according to claim 1 or 2, characterized in that, The first user characteristic factors, candidate product factors, and discount strategy factors are input into a multi-factor combination optimization algorithm. This algorithm, based on a multi-level cache and / or real-time computing engine, obtains an effective business combination suitable for the first user and its predictive value, specifically including: The multi-factor combination optimization algorithm sorts the input candidate product factors and discount strategy factors according to the selection popularity, and obtains the first effective business combination suitable for the first user and its first predictive value from the L1 hot spot cache according to the ranking. Among them, the first effective business combination is a first hot combination of candidate products with first candidate product factors and discount strategies with first discount strategy factors, and the second user characteristic factors of historical users who selected the first hot combination are more similar to the first user characteristic factors than the first preset threshold. Among them, the first candidate product factor and the first discount strategy factor are the top candidate product factors and discount strategy factors ranked according to popularity.

4. The method according to claim 3, characterized in that, The first user characteristic factors, candidate product factors, and discount strategy factors are input into a multi-factor combination optimization algorithm. This algorithm, based on a multi-level cache and / or real-time computing engine, obtains an effective business combination suitable for the first user and its predicted value. Specifically, it also includes: If the multi-factor combination optimization algorithm fails to obtain a first effective business combination and its first predicted value suitable for the first user from the L1 hot spot cache, it will obtain a second effective business combination and its second predicted value suitable for the first user from the L2 pre-calculated result cache according to the ranking. Among them, the second effective business combination is a second hot combination of candidate products with second candidate product factors and discount strategies with second discount strategy factors, and the third user characteristic factor of historical users who selected the second hot combination has a greater similarity to the first user characteristic factor than the second preset threshold. Among them, the second candidate product factor and the second preferential strategy factor are several candidate product factors and preferential strategy factors ranked according to popularity after the first candidate product factor and the first preferential strategy factor.

5. The method according to claim 4, characterized in that, The first user characteristic factors, candidate product factors, and discount strategy factors are input into a multi-factor combination optimization algorithm. This algorithm, based on a multi-level cache and / or real-time computing engine, obtains an effective business combination suitable for the first user and its predicted value. Specifically, it also includes: If the multi-factor combination optimization algorithm fails to obtain a second effective business combination and its second predicted value suitable for the first user from the L2 pre-calculation result cache, it will call the real-time calculation engine based on the first user characteristic factors, candidate product factors and preferential strategy factors to obtain a third effective business combination and its third predicted value suitable for the first user. Among them, the third effective business combination is the third combination of candidate products with the third candidate product factor and preferential strategies with the third preferential strategy factor, and the first user characteristic factor meets the business rule constraints of the third combination.

6. The method according to claim 5, characterized in that, Based on the first user characteristic factors, candidate product factors, and discount strategy factors, the real-time calculation engine is invoked to obtain the third effective business combination suitable for the first user and its third predicted value, specifically including: The business rules that call the real-time computing engine to obtain candidate product factors and discount strategy factors constitute the first business rule constraint. The real-time computing engine is invoked to obtain the second business rule constraint that the first user feature factor conforms to from the first business rule constraint; The real-time computing engine is invoked to obtain a third combination of a third candidate product factor and a third discount strategy factor from the candidate product factors and discount strategy factors according to the second business rules; The real-time computing engine is invoked to obtain the third predicted value of the third combination based on the historical user consumption data of the selected third combination. If the third predicted value already exists in the L3 intermediate result cache, it is retrieved directly; otherwise, the third predicted value is stored in the L3 intermediate result cache.

7. The method according to claim 1 or 2, characterized in that, Based on predicted value, the recommendation, acceptance, and monitoring of effective business portfolios are achieved, specifically including: Based on the predicted value, including the monthly cost of the first user after selecting an effective business combination, an effective business combination is recommended to the first user based on the monthly cost. Based on the predicted value, including the monthly cost change of the first user after selecting an effective service combination, the monthly cost change will be displayed when processing the effective service combination for the first user; The predicted value includes the change in value of the first user after selecting an effective business combination, and the value of the work of the operations staff who handle the effective business combination for the first user is evaluated based on the change in value.

8. A business portfolio value calculation application device, characterized in that, The device includes: The acquisition module is used to acquire the first user characteristic factor, candidate product factor, and discount strategy factor; The calculation module, connected to the acquisition module, is used to input the first user characteristic factors, candidate product factors and preferential strategy factors into the multi-factor combination optimization algorithm. The multi-factor combination optimization algorithm obtains the effective business combination suitable for the first user and its predictive value based on multi-level caching and / or real-time computing engine. The application module, connected to the computing module, is used to recommend, accept, and monitor effective business combinations based on predicted value. The first user characteristic factor is obtained based on the characteristics of the first user, and the effective business combination is a combination of at least one candidate product and / or a combination of at least one candidate product with at least one preferential strategy.

9. A computer device, characterized in that, The computer device includes a processor and a memory, the memory storing a computer program, and when the processor runs the computer program stored in the memory, the processor executes the business portfolio value calculation application method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the business portfolio value calculation application method as described in any one of claims 1-7.

Images