Food intake determination method and device for obtaining exercise ability, terminal and medium

Abstract

The invention provides a food intake determination method and device for acquiring athletic ability, a terminal and a medium. The method comprises the following steps: acquiring an athletic ability evaluation data set; a group of exercise ability assessment data in the exercise ability assessment data set comprises a group of user food intake data and a group of user individual feature data corresponding to the user food intake data; according to the user food intake data and the user individual feature data, constructing a human motion ability evaluation model; obtaining food variable constraint conditions and user set capability weight data; according to the food variable constraint condition, penalty term data is obtained; and generating an optimal food combination according to the food variable constraint condition, the user set capability weight data and the penalty term data. And determining the influence of different diet combination modes and nutrient intake on the exercise ability of the human body.

Description

Technical Field

[0001] This invention relates to the field of health management technology, specifically to a method, device, terminal, and medium for determining food intake to assess athletic ability. Background Technology

[0002] Research on the impact of dietary patterns and nutrient intake on the athletic performance of college students mainly focuses on cross-sectional surveys and univariate correlation analyses. Researchers often use questionnaires and 24-hour dietary recall methods to collect data, and then use descriptive statistics and simple regression analysis to explore the association between certain nutrients (such as protein, iron, and vitamins) and specific athletic performance indicators. Some studies have focused on the effects of dietary patterns such as the Mediterranean diet and high-protein diets on athletic performance, but a systematic quantitative framework is lacking.

[0003] Currently, common studies have the following significant shortcomings: (1) Static analysis dominates, neglecting the cumulative effect of nutrients in the body, metabolic dynamics and time lag characteristics, and failing to reflect the long-term dynamic effects of dietary intervention; (2) Research on nutrient interactions is weak, focusing on the independent effects of single nutrients and ignoring synergistic or antagonistic effects (such as the synergy between calcium and vitamin D, iron and vitamin C); (3) Insufficient consideration of individual differences, lacking precise analysis of individual characteristics such as age, BMI, and training level; (4) Lack of optimization decision support, existing studies mostly remain at the level of "identifying problems" and fail to provide operable optimal dietary plans; (5) Lack of integration of regional dietary culture, ignoring the impact of different regional specialty foods (such as Chongqing spicy food) on nutrient absorption and metabolism; (6) Insufficient comprehensive evaluation of multidimensional exercise ability, most studies only focus on single indicators and lack multidimensional system modeling and trade-off analysis of strength, endurance, agility and other dimensions. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention proposes a method, device, terminal, and medium for determining food intake to assess athletic performance, aiming to determine the impact of different dietary combinations and nutrient intake on human athletic performance.

[0005] In a first aspect, embodiments of this application provide a method for determining food intake to acquire athletic ability, the method comprising: Obtain a sports ability assessment dataset; the set of sports ability assessment data in the sports ability assessment dataset includes a set of user food intake data and a set of user individual characteristic data corresponding to the user food intake data; A human movement ability assessment model is constructed based on user food intake data and user individual characteristic data. Obtain data on food variable constraints and user-defined ability weights; Based on the food variable constraints, the penalty term data is obtained; The optimal food combination is generated based on food variable constraints, user-defined ability weights, and penalty data.

[0006] Secondly, embodiments of this application provide a food intake determination device, comprising: The modeling data acquisition module is used to acquire a sports ability assessment dataset; the set of sports ability assessment data in the sports ability assessment dataset includes a set of user food intake data and a set of user individual characteristic data corresponding to the user food intake data; The assessment model building module is used to build a human movement ability assessment model based on user food intake data and user individual characteristic data. The evaluation data acquisition module is used to acquire data on food variable constraints and user-defined ability weights. The penalty data determination module is used to obtain penalty item data based on the food variable constraints; The food combination determination module is used to generate the optimal food combination based on food variable constraints, user-defined ability weight data, and penalty data.

[0007] Thirdly, embodiments of this application provide a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the food intake determination method for obtaining exercise capacity as described in any of the first aspects above.

[0008] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the food intake determination method for obtaining motor ability as described in any of the first aspects above.

[0009] Fifthly, embodiments of this application provide a computer program product that, when run on a terminal device, causes the terminal device to execute the food intake determination method for obtaining motor ability as described in any of the first aspects above.

[0010] In this embodiment, a sports performance assessment dataset is acquired. This dataset includes a set of user food intake data and a set of user individual characteristic data corresponding to the food intake data. A human sports performance assessment model is constructed based on the user food intake data and user individual characteristic data. Food variable constraints and user-defined ability weights are obtained. Penalty data is obtained based on the food variable constraints. An optimal food combination is generated based on the food variable constraints, user-defined ability weights, and penalty data. The impact of different dietary combinations and nutrient intake on human sports performance is determined. Attached Figure Description

[0011] Figure 1 This is a flowchart illustrating the first embodiment of the method for determining food intake to obtain motor ability provided in this application. Figure 2 This is a schematic flowchart of a method for determining food intake to obtain motor ability according to an embodiment of this application; Figure 3 This is a schematic diagram of the data acquisition process for a method for determining exercise capacity based on food intake, provided in an embodiment of this application. Figure 4 This is a schematic diagram of a five-layer progressive relationship structure in the food intake determination method for obtaining motor ability provided in an embodiment of this application; Figure 5 This is a flowchart of the nutrient conversion process in a method for determining food intake to obtain athletic ability, provided in an embodiment of this application. Figure 6 This is a physiological metabolic flowchart of a method for determining exercise capacity based on food intake, provided in an embodiment of this application. Figure 7 This is a flowchart of the ability evaluation process in the food intake determination method for obtaining motor ability provided in an embodiment of this application; Figure 8 This is an optimization decision flowchart of a food intake determination method for obtaining motor ability provided in an embodiment of this application; Figure 9 This is a schematic diagram of the food intake determination device provided in the embodiments of this application; Figure 10 This is a schematic diagram of the structure of the terminal device provided in the embodiments of this application. Detailed Implementation

[0012] Figure 1 The illustration shows a flowchart of a first embodiment of the food intake determination method for obtaining motor ability provided in this application. This is an example and not a limitation; the method can be applied to a food intake determination device. Figure 1As shown, the method may include: S10, Obtain the sports ability assessment dataset; the set of sports ability assessment data in the sports ability assessment dataset includes a set of user food intake data and a set of user individual characteristic data corresponding to the user food intake data; To construct a human motor ability assessment model, a motor ability assessment dataset is obtained; the motor ability assessment dataset includes a set of user food intake data and a set of user individual characteristic data corresponding to the user food intake data; like Figure 2 The diagram shows the overall flowchart of the method for determining food intake to obtain exercise capacity.

[0013] like Figure 3 The diagram shows the process of collecting user food intake data and user individual characteristic data.

[0014] The steps for collecting user food intake data include: Step 111: User records a meal log: 1) The user enters food information for three meals and snacks daily through a mobile application or web interface. 2) Record content: food name, intake (g / ml), intake time, and cooking method.

[0015] Step 112: Food Coding and Standardization: 1) The system matches the food name entered by the user to a unique code in the standard food database. 2) For dialect or local specialty names (such as Chongqing "small noodles" or "wontons"), an alias mapping table is established. 3) For complex foods (such as rice bowls or hot pot), the system automatically breaks them down into their basic ingredient components. 4) For foods not included in the database, the system provides similar food recommendations or a function to manually add nutritional information.

[0016] Step 113: Data Quality Inspection: 1) Outlier Detection: Warnings are triggered when a single intake exceeds a reasonable range (e.g., 5000 grams of staple food per meal). 2) Integrity Check: User confirmation is prompted when the total daily energy intake is below 800 kcal or above 5000 kcal. 3) Logical Consistency: The rationality of food combinations is checked (e.g., contradictions in recording breakfast and late-night snack times).

[0017] Step 114: Historical Data Storage: 1) Each record includes the following fields: User ID, Date and Timestamp, Food Code, Intake, Cooking Method Code, and Meal Tag. 2) Stored using a time-series database, supporting efficient time-range queries. 3) Data retention period: At least 90 days of detailed records; data exceeding 90 days is aggregated and statistically analyzed.

[0018] The steps for collecting individual user characteristic data include: Step 121, Basic Information Collection: 1) Collection upon first use: Date of birth (to calculate age), height, weight (to calculate BMI). 2) Lifestyle habits: Smoking status, frequency of alcohol consumption, sleep duration, sedentary time. 3) Health status: Known diseases (anemia, diabetes, etc.), medication use, supplement use.

[0019] Step 122, Exercise Training Information: 1) Exercise Experience: Years of training, main sports, training frequency (times / week), training intensity. 2) Exercise Goals: Muscle gain, fat loss, endurance improvement, competition preparation, etc. 3) Special Needs: Vegetarianism, food allergies, dietary restrictions.

[0020] Steps 1, 2, and 3: Baseline Performance Tests: 1) Strength Tests: Grip strength (kg), Standing long jump (cm), Sit-ups (times / minute). 2) Endurance Tests: 800m / 1000m run time (seconds), Step test index. 3) Agility Tests: Reaction time (milliseconds), T-shaped agility test time (seconds). 4) Flexibility Tests: Sit-and-reach test (cm). 5) Balance Tests: Single-leg standing time (seconds), Single-leg standing time with eyes closed. 6) Speed ​​Tests: 50m sprint time (seconds), 10m shuttle run time.

[0021] Step 124, Regular Update Mechanism: 1) Dynamic parameters such as weight and training intensity should be updated weekly. 2) A standardized assessment of athletic ability should be conducted monthly. 3) The system will automatically remind users to update data when the cycle expires.

[0022] S20: Based on user food intake data and user individual characteristic data, construct a human movement ability assessment model; After acquiring user food intake data and user individual characteristic data, the food intake determination device constructs a human movement ability assessment model based on the user food intake data and user individual characteristic data.

[0023] The core mapping relationship of the human motor ability assessment model system is: The multifactor coupling analysis system of the relationship between overall human motor ability and multidimensional nutrient intake can be abstracted into a multi-layer nonlinear mapping: ; in, Represents a vector of athletic ability (strength, endurance, agility, flexibility, balance, speed). This represents a time-series food intake dataset. Represents an individual feature parameter vector. Represents the model parameter set, This represents a composite nonlinear mapping operator.

[0024] The human motor ability assessment model includes a five-layer progressive framework structure; In order to study the mutual coupling between various parameters in the whole system, a five-level progressive relationship structure as shown in Figure 4 is established. ; in, Indicates time Food intake matrix at time ( a kind of food, (Nutritional components) Represents nutrient vectors ( (nutrients) Represents the concentration vector of bioactive substances ( (a kind of substance) Represents the ability score vector ( (type of ability) It represents biological metabolic parameters (such as metabolic rate, absorption efficiency, etc.). Indicate personal characteristics (such as age, weight, etc.). This represents the performance model parameters (such as the weights of nutrients on capabilities). This represents the aggregated weight vector (used to calculate the overall score). It indicates the conversion of nutrients ingested from food (such as nutrient content calculation). This indicates the metabolic transformation of nutrients into bioactive substances (such as dynamic metabolic models). This represents the performance mapping from bioactive substances to capacity scores (such as linear or nonlinear regression). This represents the aggregation of ability scores to total scores (e.g., weighted summation).

[0025] As one implementation method, a human motor ability assessment model is constructed based on user food intake data and individual user characteristic data, including: A1, based on user food intake data and user individual characteristic data, obtain nutrient vector data and bioavailability data; After acquiring user food intake data and user individual characteristic data, the food intake determination device obtains nutrient vector data and bioavailability data based on the user food intake data and user individual characteristic data.

[0026] As one implementation method, such as Figure 5 As shown, based on the user's food intake data, nutrient vector data is obtained, including: B1. Construct a nutrition database; the nutrition database includes a basic database and a database of Chongqing specialty foods. The food intake layer of the human motor ability assessment model includes: Input matrix: ; Regional Adjustment (Chongqing Characteristics): ; in, Indicates the first Heavenly Intake of each type of food (grams). Indicates the size of the food database. This represents the Hadamard product (element-by-element multiplication). This represents the regional adjustment vector, reflecting the impact of cooking methods, spiciness, and other factors on actual intake.

[0027] The steps involved in building a nutrition database include: Step 201: Establishing the Basic Database: 1) Collect authoritative nutrition facts data: Chinese Food Composition Tables, USDA database, and local specialty food survey data. 2) Database fields include: food code, food name, food category, nutrient content per 100 grams (protein, fat, carbohydrates, dietary fiber, vitamins A / B1 / B2 / B6 / B12 / C / D / E / K, calcium, iron, zinc, magnesium, selenium, potassium, sodium, phosphorus, etc., 25-30 nutrients). 3) Energy density: kcal per 100 grams; 4) Cooking coefficient: nutrient retention rate when raw food is converted to cooked food.

[0028] Step 202: Expanding on Chongqing Specialty Foods: 1) Specializing in Chongqing specialty foods: hot pot base, Chongqing noodle seasoning, hot and sour noodles, spicy blood curd, spicy chicken, spring water chicken, tofu pudding, etc. 2) Recording spiciness levels (1-10) and estimated capsaicin content. 3) Investigating typical recipes and ingredient ratios (e.g., oil content in hot pot, ingredients in Chongqing noodle seasoning packets). 4) Climate adaptability labeling: The impact of Chongqing's hot and humid climate on food preservation and nutrition.

[0029] Step 203: Database Index Optimization: 1) Establish multi-level indexes: food name index, pinyin index, category index, and nutrient content range index. Regular updates: Add new foods quarterly and update nutritional data annually.

[0030] B2, Calculate nutrients from the user’s food intake data based on the nutrition database to obtain nutrient data and regional adjustment factors; The nutrient conversion layer of the human motor ability assessment model includes: nutrient conversion operators ; The nutrient database matrix includes: , Indicates the first The first of the food Content of various nutrients row vector Indicates the first of all foods Distribution of nutrients, column vector Indicates the first A vector of nutrient composition for a type of food. Represents a nutrient content matrix (database).

[0031] The quantitative models of dietary nutrients include: (1) Nutrient content matrix of single food: ; in Indicates the first Nutrient content vector of a food Indicates the first The first of the food Content of various nutrients (unit: per 100g). It indicates the total number of nutrient types (protein, fat, carbohydrates, vitamins A / B / C / D / E, calcium, iron, zinc, magnesium, selenium, etc.).

[0032] (2) Calculation of total nutritional content of dietary combinations: ; in Indicates the first Total nutrient intake vector per day Indicates the first Food in the Daily intake frequency weight (0-1). Indicates the first Food in the Actual daily intake (grams) This indicates the total number of food types.

[0033] (3) Nutrient density function: ; in Indicates the first The first of the food The density of various nutrients Indicates the first The energy density of a food (kcal / 100g), where 1000 is a standardized factor, represents the nutrient content per 1000kcal.

[0034] The steps involved in calculating nutrient data include: Step 211: Single food nutrient calculation: 1) Read the standard nutrient content vector of food i from the database. (per 100 grams). 2) Read the user's actual intake. (grams). 3) Considering the effect of cooking method k on nutrients, query the cooking coefficient matrix. 4) Calculate actual nutrient intake: Step 212: Complex Food Breakdown: 1) For complex foods, break them down into basic ingredients according to standard recipes. 2) Store the composition of common complex foods in the recipe library. 3) Calculate the nutrients for each component separately, and then summarize them.

[0035] Step 213: Daily Nutrition Summary: 1) Review all food records for the day; 2) Add up by nutrient type: total protein, total fat, total carbohydrates, etc.; 3) Calculate the energy ratio of the three macronutrients: , , 4) Calculate the energy supply ratio: .

[0036] The calculation steps for the regional moderating factor include: 1) For spicy foods, calculating the total capsaicin intake based on the spiciness level; 2) Capsaicin's effect on metabolic rate: 3) Impact of hot and humid climate: In Chongqing during the summer (May-September), the requirement for water-soluble vitamins increases by 10%. 4) Adjusted effective nutrient intake = original intake × regional adjustment coefficient.

[0037] B3. Based on the nutrient data and the user's food intake data, obtain the energy ratio data and dietary quality data; Dietary quality indicators: in Shannon diversity entropy, Represents the nutrient density function, The function representing nutrient balance Represents the nutrient intake vector ( (nutrients) This represents the overall score for dietary quality.

[0038] The calculation steps for the energy supply ratio data include: 1) Count the number of different types of food consumed that day, n; 2) Calculate the mass percentage of each type of food. 3) Shannon entropy: 4) Standardization: The value ranges from 0 to 1, with values ​​closer to 1 indicating greater diversity.

[0039] The calculation steps for dietary quality data include: Step 221: Nutrient Density Assessment: 1) Select key nutrients (protein, calcium, iron, vitamins A, C, and D). 2) Calculate the nutrient content per 1000 kcal of energy. 3) Compare with recommended intake: Achievement rate = Actual density / Recommended density. 4) Overall score: Geometric mean of the achievement rates for the six nutrients.

[0040] Step 222: Nutritional Balance Score: 1) Comparison of the energy contribution ratio of the three macronutrients with the recommended range (protein 10-15%, fat 20-30%, carbohydrates 55-65%). 2) Deviation = |Actual proportion - Recommended median| / Recommended range width. 3) Balance = 1 - Average deviation.

[0041] Step 223: Overall Dietary Quality Index: 1) ;2) Values ​​range from 0 to 100, and are graded as follows: Excellent (80+), Good (60-80), Average (40-60), Poor (<40).

[0042] B4. Nutrient vector data are obtained based on nutrient data, regional adjustment factors, energy ratio data, and dietary quality data.

[0043] A2, based on nutrient vector data and user individual characteristic data, obtain state vector data; After obtaining nutrient vector data and user individual characteristic data, the food intake determination device obtains state vector data based on the nutrient vector data and user individual characteristic data.

[0044] The physiological metabolic layer of the human motor ability assessment model includes: Bioavailability conversion: ; Time convolution: ; Kernel function matrix: ; Complete metabolic kinetics: ; in, express bioactive substances in time Concentration vector, This represents an external input or a basic generation rate vector (such as nutrient absorption, hormone secretion, etc.). Represents a diagonal or off-diagonal metabolic rate matrix, where Indicates the first The first substance to the first The degradation or transformation rate of a substance. This represents a nonlinear interaction function that describes the synergistic / antagonistic effects between substances (such as enzyme inhibition and competitive binding). The interaction term... middle, Represents the vector of available nutrients for organisms. This represents the absorption rate function (which depends on dose and individual). This represents the interaction matrix between nutrients. This represents a vector representing the cumulative concentration of nutrients in the body. Represents the metabolic decay rate matrix. This represents a non-linear interaction function.

[0045] Nutrient absorption and metabolism model (1) Nutrient bioavailability function: in Indicates the first Nutrients in Actual bioavailability per day The basal absorption coefficient (0-1) reflects the basic absorption capacity of the digestive system for that nutrient. Indicates the first The intake of various nutrients, This represents the dose-response attenuation coefficient, reflecting the decrease in absorption efficiency when there is an excess or deficiency. Indicates the first The optimal intake of each nutrient Indicates nutrients and The synergy or antagonism coefficient between them (positive values ​​are synergy, negative values ​​are antagonism).

[0046] (2) Time cumulative effect model: in Indicates the first Nutrients in The cumulative concentration in the body over the days, Indicates nutrients The metabolic decay rate (half-life related). This represents a time window function that defines the effective cumulative time range. This represents a time-lag variable.

[0047] As one implementation method, such as Figure 5 As shown, obtaining the state vector data based on nutrient vector data and user individual characteristic data includes: C1, based on nutrient vector data and user individual characteristic data, obtain bioavailability data; The calculation steps for bioavailability data include: Step 231: Determining the basal absorption rate: 1) Set a basal absorption coefficient for each nutrient. (Based on physiological literature); a) Protein: α = 0.92 (animal source), 0.85 (plant source); b) Iron: α = 0.18 (heme iron), 0.10 (non-heme iron); c) Calcium: α = 0.30 (dairy products), 0.25 (other foods); d) Fat-soluble vitamins: α = 0.70-0.85; e) Water-soluble vitamins: α = 0.85-0.95; Step 232: Dose-dependent adjustment: 1) Define the optimal intake of each nutrient. (Based on DRIs recommended intake). 2) Calculate intake deviation: 3) Absorption efficiency decline: ,in It is the attenuation coefficient.

[0048] Step 233: Nutrient Interaction Matrix: 1) Construction Interaction coefficient matrix , Indicates nutrients right Effects on absorption. 2) Synergistic effect (positive coefficient): a) Vitamin C on iron: (Increases absorption by 30%); b) Vitamin D's effect on calcium: 3) Antagonistic effect (negative coefficient): a) Phytic acid on iron and zinc: b) Calcium's effect on iron: (High calcium inhibits iron absorption); 4) Calculation of interaction effects: ; Step 234: Individual Characteristic Adjustment: 1) Age Factor: Absorption rate is reduced by 10-15% in older adults (>60 years old). 2) Health Status: Iron absorption rate is increased to twice the normal level in anemic states. 3) Training Status: Protein utilization is increased by 5-10% during high-intensity training.

[0049] Step 235: Final bioavailability: 1) ;2) : Composite value of individual regulatory factors.

[0050] C2, based on nutrient vector data and bioavailability data, nutrient concentration update data and nutrient co-product data are obtained; The calculation steps for updating nutrient concentration data include: Step 241: Setting metabolic parameters: 1) Determine the half-life for each nutrient. (Based on physiological data). a) Water-soluble vitamins (B, C): a) Fat-soluble vitamins (A, D, E, K): c) Calcium (mineral): Bone turnover d) Glycogen: 2) Calculate the attenuation rate: .

[0051] Step 242: Discrete-time update: 1) Use daily update step size . 2) No. Daily concentration update: a) Decrease in old concentration: b) Contribution of new intake: c) Update concentration: .

[0052] Step 243: Historical Window Accumulation: 1) Set the effective accumulation window W (e.g., 30 days); 2) Use exponentially weighted moving average: 3) For rapidly metabolized nutrients (such as B vitamins), W = 3-7 days; 4) For slowly metabolized nutrients (such as vitamin D), W = 30-90 days.

[0053] Step 244: Reserve State Modeling: 1) Establish three reserve forms: a) Immediately Available Pool: Nutrients in the blood circulation, fast response; b) Buffer Reserve Pool: Glycogen and fat in the liver and muscles, moderate response; c) Long-Term Reserve Pool: Calcium in bones and fat-soluble vitamins in the liver, slow response; 2) State Transition Rules: a) When intake is sufficient: Prioritize filling the immediate pool → buffer pool → long-term pool; b) When intake is insufficient: Prioritize consuming the immediate pool → buffer pool → long-term pool; c) The transition rate between different pools is different; The calculation steps for nutrient synergistic product data include: 1) Calculating synergistic products between nutrients: a) b) c) 2) Modeling using product terms and saturation functions: .

[0054] C3 inputs nutrient vector data, bioavailability data, nutrient concentration update data, and nutrient co-product data into a pre-trained physiological metabolic model, and then corrects the model using a Kalman filter algorithm to obtain state vector data. The steps for dynamically updating the state space to obtain state vector data include: Step 251: State variable initialization; 1) Upon first use, estimate the initial state based on the user's recent diet and test results. 2) If no historical data is available, use the population average as the initial state. 3) The state vector includes: a) a nutrient concentration subvector. b) Energy reserve sub-vector : Glycogen reserves, fat reserves, protein reserves; c) Internal environment sub-vectors Inflammatory markers, oxidative stress levels, and hormone level estimation; Step 252: Calculation of the state transition matrix: 1) Matrix Describing the natural evolution of the system: a) Diagonal elements: (Self-decay); b) Off-diagonal elements: Conversion relationships between nutrients; 2) Input matrix Describe the effects of external intake: a) The direct effects of a unit of food input on each state.

[0055] Step 253: Daily Status Update: 1) At the end of time t: a) Collect daily nutrient intake. b) Calculate bioavailability c) State prediction: d) Incorporating process noise: ; Step 254: Observation Update (Kalman Filter): 1) When there is actual athletic ability test data Time: a) Calculate and predict observations: b) Calculate the residuals: c) Calculate the Kalman gain d) State correction: e) This step integrates model predictions and actual measurements to improve accuracy; Step 255: Extrapolation of long-term trends: 1) For the future Predictions for the next day: a) Assumptions about future dietary patterns: historical averages, user plans, and optimization strategies can be used; b) Iterative updates: , , ..., ; Calculate prediction uncertainty: covariance matrix Follow It increases as it grows.

[0056] A3. Based on the state vector data and user individual characteristic data, a human motor ability assessment model consisting of a comprehensive ability score and a single ability score is obtained. After obtaining state vector data and user individual characteristic data, the food intake determination device generates a human motor ability assessment model consisting of a comprehensive ability score and individual ability scores based on the state vector data and user individual characteristic data.

[0057] As one implementation method, such as Figure 6 As shown, the state vector data includes nutrient concentration sub-vectors, energy reserve sub-vectors, and internal environment sub-vectors; based on the state vector data and user individual characteristic data, a human motor ability assessment model consisting of a comprehensive ability score and individual ability scores is obtained, including: D1, nutrient concentration subvector, energy reserve subvector, and internal environment subvector, yield individual ability score data; individual ability score data includes: strength ability score, endurance ability score, agility ability score, flexibility ability score, balance ability score, and speed ability score.

[0058] The capability expression layer of the human motor ability assessment model includes: Single ability generation function: ; Matrix format (6 athletic abilities): in, Represents the baseline capability vector (nature + training). Main effect weight matrix Represents a vector of nonlinear activation functions. Represents the interaction effect weight matrix. This represents the matrix vectorization operator.

[0059] Individual trait regulation: ; in Indicates the first The first nutrient for the The response function of the capability (which can be the Hill function, Sigmoid, etc.). Represents the influence matrix of individual characteristics. Represents the standardized constant of individual parameters The calculation steps for individual ability score data include: Step 261: Strength Ability Assessment: 1) Read the concentration of relevant nutrients: , , , 2) Read total energy intake: 3) Calculate the protein effect: a) Protein is the main effector; creatine has an enhancing effect but also exhibits a saturation effect; 4) Calculate the mineral effect: a) Calcium and magnesium synergistically affect muscle contraction, using geometric mean; 5) Calculate energy effect: a) Sufficient energy is fundamental, but excessive energy provides no additional benefit; 6) Weighted summaries: 7) Weight setting: (Based on literature and expert opinions).

[0060] Strength Ability Model: ; in , Indicates the first Heaven's strength and ability rating , , Represents the weighting coefficients, satisfying , This indicates the cumulative protein concentration and its main effect on muscle synthesis. This indicates the cumulative concentration of creatine (obtained through food). , This indicates the cumulative concentration of calcium and magnesium, which affects muscle contraction. Indicates total energy intake, Indicates the reference protein concentration (standardized baseline). This represents the hyperbolic tangent function, used to simulate saturation effects.

[0061] Step 262: Endurance Assessment: 1) Read relevant nutrients: ;2) Calculate the energy matrix effect: a) Carbohydrates are the primary energy source, while fats participate in long-term energy supply when carbohydrates are abundant; 3) Calculate the oxygen transport effect: a) Iron determines hemoglobin, and B12 participates in erythrocyte production; 4) Calculate hydration status: 5) Weighted summarization: 6) Weight setting: .

[0062] Endurance capacity model: ; in , This indicates the endurance rating. Indicates the weighting coefficient. This indicates the cumulative amount of carbohydrates, which are the primary source of energy. This indicates the amount of fat accumulated, which is the energy source for prolonged exercise. This indicates iron concentration, which affects hemoglobin and oxygen transport. This indicates the concentration of B vitamins, which are involved in energy metabolism. Indicators representing hydration status, This represents the carbohydrate threshold at which lipid oxidation is initiated.

[0063] Step 263: Sensitivity Assessment: 1) Reading relevant nutrients: ;2) Calculate the neural response effect: a) Stable blood glucose levels are crucial; large fluctuations slow down reaction speed; 3) Calculate neural transmission effects: a) Magnesium and potassium affect neuromuscular electrical potential conduction; 4) Calculate nervous system health: a) Omega-3 promotes nerve membrane health, but has a saturation effect; 5) Weighted summary: 6) Weight setting: ; Sensitivity model: ; in , Indicates a sensitivity rating. Indicates blood glucose concentration, which affects nerve reaction speed. This indicates potassium concentration and its influence on neuromuscular transmission. This indicates that Omega-3 fatty acids affect nervous system function, and the secondary penalty term is used to reflect the negative impact of blood sugar fluctuations on sensitivity.

[0064] Step 264: Flexibility Assessment: 1) Obtain relevant nutrients: ;2) Calculate the collagen synthesis effect: a) Collagen requires protein and vitamin C, and glycine is a key amino acid; 3) Calculate antioxidant protection: a) Vitamins C and E protect connective tissue from oxidative damage; 4) Calculate hydrated elasticity: 5) Weighted summarization: 6) Weight setting: .

[0065] Flexibility model: ; in , Indicates flexibility score, This indicates indicators related to collagen synthesis (synergistically generated by protein and vitamin C). This indicates the concentration of glycine and the main amino acids in collagen. It indicates water intake and affects the elasticity of connective tissue. Vitamins C and E have antioxidant effects and protect connective tissue.

[0066] Step 265: Balance Ability Assessment: 1) Read relevant nutrients: ;2) Calculate the bone health effect: a) Vitamin D and calcium synergistically promote bone density and affect proprioception; 3) Calculate the effects on the nervous system: a) B12 affects vestibular and nerve function; 4) Calculate the effects of trace elements: 5) Weighted summarization: 6) Weight setting: .

[0067] Balance Capacity Model: ; in , This indicates a score for balance ability. , These represent vitamin D and calcium, respectively, which affect bone health and proprioception. Vitamin B12 indicates that it affects the nervous system and vestibular function. This indicates the element zinc, which affects nerve conduction; the geometric mean reflects the synergistic effect of the two nutrients.

[0068] Step 266: Speed ​​Ability Assessment: 1) Read relevant nutrients: 2) Calculate the ATP production effect: a) ATP rapid energy supply system, vitamin B1 participates in glucose metabolism; 3) Calculate the effect of phosphocreatine: ;4) Calculate the ion conduction effect: 5) Weighted summarization: 6) Weight setting: .

[0069] Speed ​​capability model: ; in , Indicates speed ability rating. Indicators related to ATP production capacity (carbohydrates and phosphorus). This represents phosphorus, an element essential for ATP synthesis. This indicates that sodium and potassium ions affect the speed of nerve impulse conduction. It refers to vitamin B1, which participates in glucose metabolism and ATP production.

[0070] D2, based on individual user characteristic data, yields individual adjustment factors; The calculation steps for individual regulation factors include: Step 271: Construct individual feature vectors: Numerical processing: BMI classification: underweight (<18.5)=-1, normal (18.5-24)=0, overweight (24-28)=1, obese (>28)=2; Training intensity: 1-5 level scale.

[0071] Step 272: Calculate individual moderating factors: 1) Construct an influence coefficient vector for each athletic ability. (Learning through training data or expert settings); 2) Example: Adjustment of strength ability; a) Age effect: (Strength decreases by 1% for every year of age increase); b) Impact of BMI: (Moderate weight gain is beneficial for strength); c) Training effects: (Strength increases by 5% for every additional year of training); 3) Standardized individual parameters: 4) Calculate the adjustment factor: a) Use an exponential function to ensure the adjustment factor is positive.

[0072] Step 273: Applying individual adjustment: 1) Adjusted ability score: .

[0073] Individual characteristic adjustment function: ; in Indicates the first Individual characteristics for the first The influence coefficient of athletic ability Indicates the first Individual feature values, This represents the group mean of this characteristic. The standard deviation of this feature is represented by the exponential function, which ensures that the adjustment factor is always positive.

[0074] D3, based on individual ability score data and individual adjustment factors, yields a human motor ability assessment model consisting of a comprehensive ability score and individual ability scores.

[0075] The comprehensive evaluation layer of the human motor ability assessment model includes: Weighted overall score: Constrained objective function: Among the penalties: Capability weight vector Normalization ensures that different capabilities are on the same scale. Represents a constrained comprehensive utility function. This represents the penalty weight hyperparameter.

[0076] Comprehensive athletic ability evaluation: ; in This represents the overall athletic ability score. Indicates the first The basic weight of athletic ability Indicates the first The assessment of athletic abilities (strength, endurance, agility, flexibility, balance, speed). Indicates individual characteristic moderating factors, For individual parameters such as strength (age, BMI, training level, etc.).

[0077] The calculation steps for the comprehensive ability score include: Step 281, Normalization Process: 1) Determine the theoretical scope of each capability. (Through statistical analysis of a large number of samples); 2) Normalization: 3) After normalization, the value is 0-1, which facilitates comparison between different abilities.

[0078] Step 282, Weight Setting: 1) Default balanced weight: μ = [1 / 6, 1 / 6, 1 / 6, 1 / 6, 1 / 6, 1 / 6]; 2) User-defined weights are supported: such as those for sprint races. The rest are all 0.075; 3) Inspection constraints: .

[0079] Step 283, calculate the overall score: ; Step 284, Generate radar image data: Output a six-dimensional vector: Used to visualize and display capability structures.

[0080] The steps involved in parameter learning and model training for human motor ability assessment models include: E1, Data Preparation: Step 601, Data Collection: 1) Collect training data from multiple sources: Clinical research data: results of nutritional intervention experiments; Athlete monitoring data: dietary diaries + ability tests; College student survey data: questionnaires + physical fitness tests; 2) Data scale: at least 500-1000 samples (individuals), with each sample recorded for at least 30 days.

[0081] Step 602, Data Preprocessing: 1) Missing Value Handling: a) Food Data: Use forward imputation or interpolation; b) Ability Tests: Use linear interpolation or LOCF (Last Observation Forward Imputation); 2) Outlier Handling: a) Identify outliers using the 3σ principle or IQR method; b) Mark outliers but do not delete them, assign them low weights; 3) Data Standardization: a) Nutrient Data: Z-score standardization; b) Ability Scores: Min-Max normalization to 0-100.

[0082] Step 603, Feature Engineering: 1) Constructing Derivative Features: a) Nutrient Ratios: such as calcium-phosphorus ratio, Comparison; b) Sliding window characteristics: average intake over the past 7 days; c) Trend characteristics: trend of increase or decrease in intake; 2) Time characteristics: a) Day of the week (weekend vs. weekday); b) Season (affects food types).

[0083] Step 604, Dataset Partitioning: 1) Training set: 70% (for parameter learning); 2) Validation set: 15% (for hyperparameter tuning); 3) Test set: 15% (for final evaluation); 4) Time series data Note: Partition by individual to avoid data from the same individual being scattered across different sets; E2, then perform parameter initialization. Step 611, Physical constraint initialization: 1) Absorption rate 1) Refer to the literature to set initial values; 2) Half-life : Based on physiological knowledge; 3) Weighting Uniform distribution; Step 612, Initialization of empirical rules: 1) Nutrient-ability correlation weights; a) Protein to strength: high weight; b) Iron to endurance: high weight; c) Others are set according to nutritional common sense; 2) Individual characteristic coefficients a) Estimated based on effect size from the literature.

[0084] Step 613, random perturbation: 1) Add small noise (e.g., ±10%) to the initial value; 2) Avoid getting trapped in local optima.

[0085] E3, Parameter Optimization Implementation Step 621, Define the loss function: 1) Prediction error: ;2) Regularization term: 3) Physical constraint penalties: (Absorption rate cannot be negative); 4) Total loss: ; Step 622, select the optimization algorithm: 1) Gradient descent class: a) Adam: adaptive learning rate, suitable for large-scale parameters; b) L-BFGS: quasi-Newton method, suitable for medium-scale; 2) Bayesian optimization: a) optimize hyperparameters (such as regularization coefficient λ); b) use Gaussian process to model the objective function; 3) Evolutionary algorithm: a) suitable for non-convex and non-differentiable cases.

[0086] Step 623, Backpropagation: 1) Calculate the gradient: ;2) The chain rule propagates backward layer by layer;3) For discrete variables or non-differentiable elements, use numerical gradient.

[0087] Step 624, Parameter Update: 1) Update the formula (taking Adam as an example): ; ; ;2) Learning rate scheduling: gradually decrease from 0.01 to 0.0001;3) Early stopping mechanism: stop if the validation set loss does not decrease for 10 consecutive rounds.

[0088] Step 625, Model Validation: 1) Evaluate on the validation set: a) (Root Mean Square Error); b) (mean absolute error); c) (Coefficient of determination); 2) Cross-validation: a) a) Cross-validation with folded cross-validation (K=5 or 10); b) Time series cross-validation (rolling window).

[0089] Step 626, Hyperparameter Tuning: 1) Parameter to be tuned: Regularization coefficient Learning rate 1) Network structure (e.g., number of hidden layers); 2) Methods: a) Grid search: exhaustive combination (small scale); b) Random search: random sampling of parameters (medium scale); c) Bayesian optimization: intelligent exploration (large scale). E4, Model Integration Step 631, train multiple sub-models: 1) Different initializations: train multiple times with different random seeds; 2) Different data: train using Bootstrap sampling. Step 632, Integration Strategy: 1) Averaging Method: 2) Weighted average: ( (Determined based on validation set performance); 3) Stacking: Using sub-model predictions as features; training meta-learners for fusion; Step 633, Uncertainty Quantification: 1) Prediction Variance: Confidence interval: .

[0090] Dietary pattern vector representation: ; in Indicates time period Dietary pattern feature vector within, , , These represent the energy provided by the three major nutrients. It indicates the intake proportion and frequency of a certain type of food. The energy percentage reflects the macro-nutrient structure, and the food proportion reflects the dietary pattern type.

[0091] Dietary pattern cluster distance: ; in This represents the weighted Euclidean distance between two dietary patterns. Indicates the first The importance weights of each feature are used for cluster analysis to identify similar dietary patterns.

[0092] Dietary diversity index: ; in , Shannon entropy represents dietary diversity. Indicates the first The weighting of each food group in total intake. Indicates the first The higher the frequency and variety of food intake, the better. The larger the value.

[0093] Multivariate nonlinear regression model ; in Indicates lag The time after athletic ability, Indicates baseline athletic ability (innate and long-term training factors). Indicates the first The first nutrient for the The main effect coefficient of athletic ability, Represents nonlinear transformation functions (such as logarithmic, square root, and sigmoid functions). Indicates nutrients and The interaction effect coefficient, Represents the random error term. .

[0094] Dose-response curve: ; in Indicates the first The dosage of each nutrient is At the time of the first The response of athletic ability This indicates the maximum response value (upper limit effect). It represents the half-maximal effective concentration, reflecting the "potency" of a nutrient. The Hill coefficient represents the steepness (synergism) of the dose-response curve.

[0095] Time lag effect model: ; Kernel function , This represents the Gamma distribution kernel function, which describes the decay of the effect of nutrient intake over time. This represents the shape parameter, which controls the timing of peak occurrence. This represents a scale parameter that controls the duration of the effect. This represents the Gamma function, and its integral reflects the cumulative effect of historical nutrient intake.

[0096] To simultaneously optimize multiple conflicting performance metrics and output the optimal dietary plan under budget and nutritional constraints, the multi-objective optimization objective function includes: Constraints: Total food quantity constraint: Nutrient range constraints: Energy constraints: Non-negativity constraint: Overcomputation constraint: .

[0097] in This represents the decision variable, a vector of the intake of each food item. This indicates the daily total food intake limit. Indicates the first Recommended intake range for each nutrient. This represents the total cost of the meal. This indicates the budget ceiling.

[0098] Since it is impossible to improve all objectives simultaneously, we employ the Pareto optimality approach to provide multiple trade-offs. Pareto front set: in This represents the Pareto optimal solution set, meaning there are no other solutions that can improve other capabilities without reducing one capability.

[0099] To facilitate the solution, we transformed the multi-objective optimization problem into a single-objective problem and introduced a penalty term to ensure nutrient balance and cost control.

[0100] The normalization term transforms the different dimensions of motor ability into a 0-1 interval. This indicates a penalty for nutrient imbalance. This indicates a penalty for excessively high costs. This represents the penalty weighting coefficient.

[0101] Nutritional Imbalance Penalty Function ; The secondary penalty ensures that a greater penalty is imposed when nutrients are severely exceeded or insufficient, and the penalty is only activated when they exceed or fall below the recommended range.

[0102] To capture the evolutionary patterns of hidden variables such as nutrient reserves and metabolic status in the body, a dynamic monitoring and prediction model was constructed.

[0103] State-space model ; in This represents the hidden state vector (nutrient reserves, metabolic state, etc.). This represents the observable motion capability vector. Represents the nutrient input vector. The state transition matrix represents the natural evolution of a physiological system. This represents the input effect matrix, describing how nutrient intake changes the state. The observation matrix describes how the hidden state manifests as mobility. , These represent process noise and observation noise, respectively.

[0104] Kalman filter predictions include: state estimation: Kalman gain: ; in Indicates based on the observation of the first The optimal state estimate for the day This represents the Kalman gain, balanced model predictions, and actual observations. This represents the prediction error covariance matrix. This represents the observation noise covariance matrix.

[0105] Long-term trend forecast ; in Indicates the predicted time span (number of days). Reflecting the natural evolution of the current state, It reflects the cumulative impact of future nutrition intake plans.

[0106] To accurately identify which factors have the greatest impact on different sports outcomes, a sensitivity analysis method was introduced.

[0107] Partial derivative sensitivity ; in Indicates the first The first nutrient for the The sensitivity of a nutrient to athletic ability is such that the higher the value, the more significant the effect of that nutrient on athletic ability.

[0108] Elasticity coefficient ; in The elasticity coefficient represents the percentage change in athletic performance resulting from a percentage change in nutrient intake. It is a dimensionless index, facilitating comparisons between different nutrients. This indicates high flexibility (sensitivity). This indicates low elasticity (insensitivity).

[0109] Regional food nutritional regulators ; in This indicates the effective nutrient content of food in the Chongqing area. Indicates standard nutrient content, This indicates the effect coefficient of capsaicin (such as promoting metabolism and affecting absorption). Indicates the first The spiciness index (0-1) of a food. Indicates the influence coefficient of hot and humid climate. This indicates the seasonal humidity level.

[0110] For common Chongqing foods (hot pot, Chongqing noodles, hot and sour noodles, spicy blood curd, spicy chicken, etc.), additional considerations should be made: ; in Indicates the first Total daily capsaicin intake Indicates the first The capsaicin content of a certain food.

[0111] Effects of capsaicin on metabolism: in This indicates the basal metabolic rate. Capsaicin can increase the metabolic rate, but the effect is saturated.

[0112] We fit all model parameters using real data and utilize R... 2 RMSE and MAPE are used to assess the accuracy of predictions to ensure the reliability of the model.

[0113] Parameter estimation ; Log-likelihood: ; in It represents the set of all parameters of the model (weights, coefficients, thresholds, etc.). Represents the observation dataset, This represents the likelihood function, and the optimal parameters are determined by fitting actual data.

[0114] Model validation metrics include the coefficient of determination: Root mean square error: Mean absolute percentage error: in This represents the model's predicted value. This represents the mean of the observed values. A value close to 1 indicates a good fit. and The smaller the value, the smaller the prediction error.

[0115] S30: Obtain food variable constraints and user-defined capability weight data.

[0116] To determine the impact of different dietary combinations and nutrient intake on human exercise capacity, data on food variable constraints and user-defined ability weights were obtained.

[0117] The food variable constraints include: Step 301, Define decision variables: 1) Decision variables: ,express 1) Intake of each type of food (grams); 2) Variable dimensions: Typically, this includes 50-200 common foods; 3) Range of variables: (Reasonable upper limit for each type of food); Step 302, set constraints: 1) Nutrient constraints: a) For each nutrient j: b) Set boundaries with reference to DRIs (Dietary Reference Intakes); 2) Energy constraints: a) Calculated based on user weight and activity level: 3) Total food intake constraints: a) (Stomach capacity limitation, usually 2-3 kg / day); 4) Budget constraints: Example: A college student's average daily meal cost is 30-50 yuan; Non-negative constraint: .

[0118] Constructing the objective function, user-defined capability weight data includes: 1) Multi-objective vector: ;2) Each goal is Functions: 3) Calculate using the aforementioned model chain: .

[0119] S40, based on the food variable constraints, obtain the penalty term data; After obtaining the food variable constraints, the food intake determination device obtains penalty term data based on the food variable constraints. The calculation steps for the penalty data include: 1) Nutritional imbalance penalty: Both exceeding and falling short are penalized, with quadratic functions exhibiting severe deviations. 2) Cost penalty: Activated only when budget is exceeded. 3) Diversity penalty: (Negative entropy, encouraging diversity), avoid eating only a few types of food.

[0120] S50 generates the optimal food combination based on food variable constraints, user-defined ability weight data, and penalty data.

[0121] like Figure 7 As shown, after obtaining the food variable constraints, user-defined ability weight data, and penalty data, the food intake determination device generates the optimal food combination based on these constraints.

[0122] As one implementation method, generating the optimal food combination based on food variable constraints, user-defined ability weight data, and penalty data includes: iteratively optimizing the food variable constraints, user-defined ability weight data, and penalty data using a single-objective optimization algorithm to generate the optimal food combination.

[0123] The optimal food combination is generated by iteratively optimizing the food variable constraints, user-defined ability weights, and penalty data using a single-objective optimization algorithm. Step 501, Weighted Transformation of Objective Function: 1) User sets weights 2) Single objective function: 3) Normalization ensures that different capabilities are on the same scale; Step 502, select the optimization algorithm: 1) For small-scale problems (n<100): use SLSQP (Sequential Least Squares Programming) or interior point method; 2) For large-scale problems (n>100): use heuristic algorithms such as genetic algorithm, particle swarm optimization, simulated annealing, etc.; 3) Gradient calculation: If the model is differentiable, calculate... Use gradient-based optimization; otherwise, use gradient-free methods or numerical gradients.

[0124] Step 503, Initial solution generation: 1) Method 1: Use the user's historical average diet as the starting point; 2) Method 2: Generate a random feasible solution; 3) Method 3: Use heuristic rules (such as the minimum cost solution that meets energy requirements).

[0125] Step 504, Iterative Optimization: 1) Iterative Steps: Calculate the objective function value and gradient of the current solution; determine the search direction; determine the step size using line search; update the decision variables: Check convergence conditions: 2) Maximum number of iterations: 1000-5000; Step 505, Feasibility Repair: 1) If an infeasible solution (constraint violation) is generated during the optimization process: a) Projection method: Project the solution to the boundary of the feasible region; b) Penalty method: Increase the penalty coefficient for constraint violation; c) Repair heuristic: Adjust the variables to satisfy the constraints.

[0126] Step 506, output the optimal solution: 1) Optimal food combination: 2) Filter non-zero items: only keep 3) Generate readable recipes: a) Allocate by meal (30% for breakfast, 40% for lunch, 25% for dinner, and 5% for snacks); b) Convert into specific dishes (e.g., "200g rice + 100g chicken breast + 150g broccoli").

[0127] As one implementation method, generating the optimal food combination based on food variable constraints, user-defined ability weight data, and penalty data includes: F1 generates a set of random feasible dietary plans based on food variable constraints and penalty data; F2 processes the constraints of the material variables, user-defined capability weights, penalty data, and the set of random feasible dietary schemes using non-dominated sorting methods, crowding calculation methods, genetic operation methods, environmental selection methods, decision support methods, and time series optimization methods to generate the optimal food combination.

[0128] The steps to obtain the optimal food combination through multi-objective optimization include: Step 511, Initial Population Generation: 1) Generation A random feasible solution ;2) Each individual represents a dietary program;3) Ensure the diversity of the initial population (covering different areas of the target space); Step 512, Non-dominated sorting: 1) For each solution Compare its relationship with all other solutions: a) If Make b) Solutions not dominated by any solution constitute the first frontier (Pareto optimality); c) After removing the first frontier, the remaining solutions that are not dominated constitute the second frontier; d) and so on. Step 513, Crowding Calculation: 1) For solutions to the same frontier, calculate the crowding distance: a) For each target , will unlock Sort; b) Boundary solution crowding distance = ∞; c) Intermediate solution: 2) A large crowding distance indicates that the solutions are more dispersed in the target space, and these solutions should be retained first. Step 514, Genetic Operations: 1) Selection: Tournament selection, prioritizing individuals with high frontier rank and large crowding distance; 2) Crossover: a) Select two parents b) Arithmetic crossover: c) or SBX (simulated binary crossover); 3) Mutation: a) Polynomial mutation: b) ~A certain probability distribution is used to control the mutation amplitude; 4) Constraint handling: After crossover mutation, check and repair constraint violations; Step 515, Environmental Selection: 1) Merge parent and offspring populations; 2) Select by non-dominant rank and crowding distance. Individuals enter the next generation; 3) Prioritize the retention of those with higher frontier levels, and retain those with larger crowding distances within the same frontier.

[0129] Step 516, Termination and Output: 1) Termination condition: Reaching the maximum number of generations (e.g., 100-500 generations) or frontier stability (no improvement for multiple consecutive generations); 2) Output Pareto solution set: 10-50 non-dominated solutions; 3) Display the trade-off curve: Two-dimensional: such as the Pareto frontier of strength vs. endurance; Multi-dimensional: using a parallel coordinate graph to display the 6-dimensional objective.

[0130] Step 517, Decision Support: 1) Users select from the Pareto solution set: provide an interactive interface to adjust preferences and see the changes in the solution; or use decision rules: such as the solution closest to the ideal point (the virtual point where all objectives are optimal); or the knee method: the point with the largest curvature of the frontier curve (the best trade-off). The steps involved in time series optimization (i.e., dynamic programming) include: Step 521, Define the time window: 1) Optimization period T: 7 days (one week) or 30 days (one month); 2) Daily decisions: 3) Total decision variables: , ; Step 522, State space definition: 1) State variables: (Nutrient concentration, energy reserves, internal environment); 2) State transition: (Determined by the metabolism module); 3) Initial state: Determined by current physical condition; Step 523, Multi-stage objective function: 1) Cumulative objective: ;2) The discount factor (e.g., 0.95) has a higher weighting for recent returns; 3) Future states depend on past decisions: ; Step 524, Decomposition Method: 1) Method 1: Rolling Time Domain Optimization (MPC Idea); a) Optimize only the future each time. Heaven (as) b) After executing the plan for day 1, roll it over to the next day for re-optimization; c) Advantages: Reduced computational complexity, adaptable to dynamic changes; 2) Method 2: Complete optimization: a) Optimize all at once a) Using dynamic programming or gradient descent; c) Disadvantages: high computational cost, but the solution is better; Step 525, Periodic Constraints: 1) Daily Constraints: Meet nutritional and budget constraints every day; 2) Periodic Constraints: Such as at least 3 high-protein meals per week and at least 5 different types of meat per week; 3) Inventory Constraints: Such as fresh vegetables need to be consumed within 2 days.

[0131] Step 526, Uncertainty handling: 1) Robust optimization: Consider the uncertainty of nutrient absorption; 2) Stochastic optimization: Simulate multiple possible scenarios and optimize the expected value; 3) Sensitivity analysis: Evaluate the stability of the scheme to parameter changes.

[0132] In summary, this application obtains a sports performance assessment dataset; the dataset includes a set of user food intake data and a set of user individual characteristic data corresponding to the food intake data; based on the user food intake data and user individual characteristic data, a human sports performance assessment model is constructed; food variable constraints and user-defined performance weights are obtained; penalty data is obtained based on the food variable constraints; and an optimal food combination is generated based on the food variable constraints, user-defined performance weights, and penalty data. This determines the impact of different dietary combinations and nutrient intake on human sports performance.

[0133] For those consistent with the above, please refer to Figure 9 , Figure 9 This application provides a schematic diagram of the structure of a food intake determination device. Figure 9 As shown, the device includes: The modeling data acquisition module 901 is used to acquire a sports ability assessment dataset; the set of sports ability assessment data in the sports ability assessment dataset includes a set of user food intake data and a set of user individual characteristic data corresponding to the user food intake data. The assessment model building module 902 is used to build a human movement ability assessment model based on user food intake data and user individual characteristic data. The evaluation data acquisition module 903 is used to acquire data on food variable constraints and user-defined ability weights. The penalty data determination module 904 is used to obtain penalty item data based on the food variable constraints; The food combination determination module 905 is used to generate the optimal food combination based on food variable constraints, user-defined ability weight data, and penalty data.

[0134] This application also provides a terminal device, which includes: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor. When the processor executes the computer program, it implements the steps in the method embodiment for determining exercise capacity.

[0135] This application also provides a computer storage medium storing a computer program for electronic data interchange, which causes a computer to perform some or all of the steps of any of the methods for determining exercise capacity as described in the above method embodiments.

[0136] This application also provides a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program that causes a computer to perform some or all of the steps of any of the methods for determining exercise capacity as described in the above method embodiments.

Claims

1. A method of determining food intake for obtaining athletic ability, characterized by, The method includes: Obtain a sports ability assessment dataset; the set of sports ability assessment data in the sports ability assessment dataset includes a set of user food intake data and a set of user individual characteristic data corresponding to the user food intake data; A human movement ability assessment model is constructed based on user food intake data and user individual characteristic data. Obtain data on food variable constraints and user-defined ability weights; Based on the food variable constraints, the penalty term data is obtained; The optimal food combination is generated based on food variable constraints, user-defined ability weights, and penalty data.

2. The method according to claim 1, wherein The process of constructing a human motor ability assessment model based on user food intake data and individual user characteristic data includes: Nutrient vector data is obtained based on the user's food intake data; Based on nutrient vector data and user individual characteristic data, state vector data is obtained; Based on the state vector data and individual user characteristic data, a human motor ability assessment model consisting of a comprehensive ability score and individual ability scores is obtained.

3. The method according to claim 2, wherein Based on user food intake data, nutrient vector data is obtained, including: Construct a nutrition database; the nutrition database includes a basic database and a database of Chongqing specialty foods. Nutrient calculations are performed on the user's food intake data based on the nutrition database to obtain nutrient data and regional adjustment factors. Based on the nutrient data and the user's food intake data, energy ratio data and dietary quality data are obtained; Nutrient vector data are obtained based on nutrient data, regional adjustment factors, energy ratio data, and dietary quality data.

4. The method for determining food intake to acquire athletic ability according to claim 2, characterized in that, The process of obtaining state vector data based on nutrient vector data and individual user characteristic data includes: Bioavailability data is obtained based on nutrient vector data and individual user characteristic data; Nutrient vector data and bioavailability data are processed to obtain nutrient concentration update data and nutrient synergistic product data. Nutrient vector data, bioavailability data, nutrient concentration update data, and nutrient co-product data are input into a pre-trained physiological metabolic model, and the model is corrected using the Kalman filter algorithm to obtain state vector data.

5. The method for determining food intake to acquire athletic ability according to claim 2, characterized in that, The state vector data includes nutrient concentration sub-vectors, energy reserve sub-vectors, and internal environment sub-vectors; based on the state vector data and user individual characteristic data, a human motor ability assessment model consisting of a comprehensive ability score and individual ability scores is obtained, including: Nutrient concentration subvector, energy reserve subvector, and internal environment subvector are used to obtain individual ability score data; Individual adjustment factors are obtained based on individual user characteristic data; Based on individual ability score data and individual adjustment factors, a human motor ability assessment model consisting of comprehensive ability score and individual ability score is obtained.

6. The method for determining food intake to acquire motor ability according to any one of claims 1 to 5, characterized in that, The process of generating the optimal food combination based on food variable constraints, user-defined ability weight data, and penalty data includes: The optimal food combination is generated by iteratively optimizing the constraints of food variables, user-defined ability weights, and penalty terms using a single-objective optimization algorithm.

7. The method for determining food intake to acquire motor ability according to any one of claims 1 to 5, characterized in that, The process of generating the optimal food combination based on food variable constraints, user-defined ability weight data, and penalty data includes: Generate a set of random feasible dietary plans based on food variable constraints and penalty data; The optimal food combination is generated by processing the constraints of the material variables, user-defined capability weights, penalty terms, and a set of random feasible dietary schemes using non-dominated sorting methods, crowding calculation methods, genetic operation methods, environmental selection methods, decision support methods, and time series optimization methods.

8. A food intake determination device, characterized in that, include: The modeling data acquisition module is used to acquire a sports ability assessment dataset; the set of sports ability assessment data in the sports ability assessment dataset includes a set of user food intake data and a set of user individual characteristic data corresponding to the user food intake data; The assessment model building module is used to build a human movement ability assessment model based on user food intake data and user individual characteristic data. The evaluation data acquisition module is used to acquire data on food variable constraints and user-defined ability weights. The penalty data determination module is used to obtain penalty item data based on the food variable constraints; The food combination determination module is used to generate the optimal food combination based on food variable constraints, user-defined ability weight data, and penalty data.

9. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method for determining food intake to obtain motor ability as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the method for determining food intake to obtain motor ability as described in any one of claims 1 to 7.

Images