Safe inventory configuration and replenishment strategy recommendation method and device

Abstract

The invention provides a method and a device for configuring safety stock and replenishment strategies in a supply chain system, and the method comprises the steps: designing time sequence characteristic parameters according to the analysis of demand characteristics; according to the intermittency, the absolute variability and the relative variability, the demand distribution type is judged; recommending a replenishment strategy according to the demand type of the demand distribution; deducing a safe inventory calculation method under non-normal distribution by taking a demand distribution judgment result as an initial judgment basis of safe inventory calculation and combining statistics and actual data; the device comprises modules determined according to the method, namely a demand time sequence parameter design module, a demand distribution judgment module, a replenishment strategy recommendation module, a safety stock calculation module and a safety stock correction module. According to the method, the safety stock and replenishment strategy of each facility-product node in the supply chain system under the established service level is expected to be optimally configured, and the method plays a positive role in improving the performance of the supply chain.

Description

Technical Field

[0001] The present invention relates to the technical field of safety stock control in supply chain management, and in particular to a safety stock configuration and replenishment strategy recommendation method and device for safety stock quantity determination technology under a given service level. Background Art

[0002] In supply chain management, the basic idea of ​​safety stock control is to reduce safety stock on the basis of effectively dealing with uncertainty and meeting a certain level of customer service. The control methods for reducing specific inventory include but are not limited to the following:

[0003] 1. Quantitative method: Determine the optimal level of inventory based on historical data and statistical analysis.

[0004] 2. Risk assessment-based approach: Assess various risks in the supply chain and adjust the safety stock accordingly.

[0005] 3. Method based on demand forecasting: dynamically adjust safety stock by predicting changes in market demand.

[0006] 4. Regular replenishment strategy, determine the replenishment time interval and the quantity of each replenishment, and determine the maximum inventory and safety stock based on the demand rate and replenishment lead time.

[0007] The above methods have high requirements on the quality and processing capabilities of historical data and demand forecast data. If the data is inaccurate or the data processing capabilities are insufficient, the calculation results of safety stock may be unreliable. At the same time, some statistical calculation methods, such as the most commonly used calculation formula: safety stock = service level coefficient * standard deviation of daily demand * square root of lead time, assume that the demand distribution conforms to the normal distribution, which is not always true in practice. Summary of the invention

[0008] In order to overcome the above-mentioned deficiencies of the prior art, the present invention provides a method and device for configuring safety inventory and replenishment strategy in a supply chain system when demand is non-normally distributed.

[0009] A safety stock configuration and replenishment strategy recommendation method, comprising:

[0010] Design time series characteristic parameter step, based on the analysis of demand characteristics, design parameters to describe the demand time series characteristics,

[0011] Determine the demand distribution type step, determine the demand distribution type based on intermittency, absolute variability, and relative variability;

[0012] Recommend replenishment strategy steps, and recommend replenishment strategies based on the demand type of demand distribution;

[0013] The safety stock calculation steps are based on the results of demand distribution judgment as the initial judgment basis for safety stock calculation. Combining statistics and actual data, the safety stock calculation method under non-normal distribution is derived.

[0014] Preferably, the parameters designed to describe the characteristics of the demand time series include intermittency, absolute variability, relative variability, and demand type.

[0015] Preferably, the step of determining the demand distribution type embeds a built-in algorithm, namely, a method for calculating intermittency, absolute variability, relative variability, and demand type.

[0016] Preferably, the step of designing characteristic parameters of time series excludes extreme demand time series that are not worthy of calculating safety stocks or recommending replenishment strategies by setting limit values ​​for the number of non-zero demand periods and the non-zero demand mean.

[0017] Preferably, the judgment demand distribution type is normal distribution, gamma distribution or negative binomial distribution.

[0018] Preferably, the safety stock calculation step embeds the safety stock calculation method under normal distribution, gamma distribution and negative binomial distribution, that is, the quantile function PPF is calculated as the basic value of the safety stock.

[0019] Preferably, the safety stock calculation method under the gamma distribution and negative binomial distribution is:

[0020] For the gamma distribution, calculate the shape parameter α and scale parameter β and set the service level service level , according to the quantile function PPF, a basic value of safety stock is obtained, that is, PPF (α, β, service level );

[0021] For the negative binomial distribution, calculate the success probability p and the number of successes r and set the service level service level , according to the quantile function PPF, a basic value of safety stock is obtained, that is, PPF (r, p, service level ).

[0022] Preferably, it also includes an adjustment and correction step, in which professional simulation software is used to set the demand time series, replenishment strategy, and safety stock, wherein the replenishment strategy and safety stock, that is, the recommended replenishment strategy and safety stock value are calculated through the demand time series through the replenishment recommendation strategy step and the safety stock calculation step, to test the effectiveness of the safety stock setting.

[0023] Preferably, it also includes recording the frequency and severity of out-of-stock and overstocking after the simulation software is finished. If the product has valuable information, the out-of-stock cost and overstocking cost can also be calculated. If the safety stock calculation device determines that the demand distribution of a facility-product node is a non-normal distribution, and there is a high probability of serious out-of-stock or overstocking, data analysis is performed to correct the safety stock calculation method.

[0024] The device of the safety inventory configuration and replenishment strategy recommendation method described in the present invention comprises:

[0025] The demand distribution determination module designs parameters such as intermittency, absolute variability, relative variability, and demand type based on the analysis of demand characteristics;

[0026] Demand distribution type determination module, which determines the demand distribution type based on intermittency, absolute variability, and relative variability;

[0027] The replenishment strategy recommendation module recommends replenishment strategies based on demand distribution type;

[0028] The safety stock calculation module uses the result of demand distribution as the initial judgment basis for safety stock calculation, and combines statistics and actual data to derive the safety stock calculation method under non-normal distribution;

[0029] The safety stock correction module uses simulation methods to determine the applicability and adjustment direction of the safety stock calculation method based on out-of-stock and over-stock situations.

[0030] The present invention has the following positive effects:

[0031] 1. Improve the accuracy of demand distribution analysis and safety stock calculation, reduce overstocking and out-of-stock situations, and reduce inventory costs;

[0032] 2. Automatically recommend replenishment strategies to provide a basis for purchasing plans. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a step diagram of a safety stock configuration and replenishment strategy recommendation method described in the present invention.

[0034] Figure 2 It is a flow chart of a safety stock configuration and replenishment strategy recommendation method described in the present invention.

[0035] Figure 3 It is a demand transmission diagram of a safety inventory configuration and replenishment strategy recommendation method described in the present invention. DETAILED DESCRIPTION

[0036] The following is a detailed description of the example of determining the demand type and calculating the safety stock according to the demand type of the present invention in conjunction with the accompanying drawings. For ease of understanding, those skilled in the art should recognize that the embodiments described below are only exemplary and various changes and optimizations are actually required.

[0037] Figure 1 It is a step diagram of a safety stock configuration and replenishment strategy recommendation method described in the present invention. Figure 2 It is a flow chart of a safety stock configuration and replenishment strategy recommendation method described in the present invention. Figure 3 It is a demand transmission diagram of a safety inventory configuration and replenishment strategy recommendation method described in the present invention.

[0038] In the attached figure Figure 2 The implementation steps of the present invention are demonstrated in detail, as well as the specific usage of intermittency, absolute variability, relative variability, and demand type in determining demand distribution and recommending replenishment strategies, presenting the architecture of the present invention clearly and logically.

[0039] In the attached figure Figure 3 It mainly shows that demand transmission in the supply chain is carried out through nodes and feasible paths between nodes. The so-called node is a demand pair consisting of a customer and a product. The customer needs to buy the same product from a facility, which constitutes a path, and a facility and the product constitute another demand pair. This idea of ​​demand transmission is also reflected in the demand distribution, that is, the demand distribution of the customer-product node will also be transmitted to the facility-product node that supplies it.

[0040] A safety inventory configuration and replenishment strategy recommendation method according to the present invention comprises:

[0041] Step 1: Analysis of demand characteristics;

[0042] Step 2: Recommend replenishment strategy based on demand type;

[0043] Step 3, safety stock calculation, wherein a calculation method for non-normal demand distribution is proposed by the present invention;

[0044] Step 4: Adjust and calibrate the safety stock calculation method based on the simulation method.

[0045] Among them, step one includes:

[0046] (1) Generate demand time series by aggregating customer demands.

[0047] The demand time series refers to a sequence of demand values ​​arranged in the order of their occurrence. First, determine the time granularity, such as "day", "week", "month", etc., divide the unit time interval according to the time granularity, and sum up the demand according to the unit time interval. There is only one demand value in a time interval. Sort by time interval to get the demand time series, which can be expressed as D(t), where t represents the unit time interval. Demand is transmitted according to the channel and procurement rules to generate the demand time series for each facility-product combination.

[0048] The historical input data should be as complete as possible and undergo preliminary data cleaning, such as processing missing or outliers. In order to verify and adjust the formula, try to prepare multiple sets of data. Each set of data has two main dimensions, namely the time of demand occurrence and the demand value.

[0049] (2) Conduct demand analysis based on this time series and design parameters to describe demand characteristics.

[0050] First, calculate the non-zero demand, which is the value of demand that is not 0 per unit time in the demand sequence. The two demand sequences may have the same demand mean and demand standard deviation, but if we observe the actual demand fluctuations in the demand time series, we will find that the two are inconsistent. Introducing non-zero demand can avoid the rough characterization of demand distribution caused by only using the demand mean and demand standard deviation.

[0051] Based on non-zero demand, three parameters can be derived: the number of non-zero demand periods, the mean of non-zero demand, and the variance of non-zero demand (standard deviation of non-zero demand).

[0052] The number of non-zero demand periods can be obtained by directly counting the number of unit time intervals with non-zero demand values, which can be recorded as P NZ ; Screen out non-zero demand D(t) NZ ,

[0053] Non-zero demand mean

[0054] Non-zero demand variance Non-zero demand standard deviation

[0055] (3) Based on the number of non-zero demand periods, the non-zero demand mean, the non-zero demand standard deviation and the empirical formula, the four demand indicators of the facility-product combination are calculated, namely, intermittency, absolute variability, relative variability and demand type.

[0056] Demand type can more accurately describe the demand distribution characteristics of the facility-product combination.

[0057] Safety stock calculation methods can be adopted under different demand distributions.

[0058] Intermittency is reflected as the ratio of the total number of periods in the time series to the number of non-zero demand periods in the demand series. The larger the ratio, the greater the intermittent nature of the demand.

[0059] Absolute variability is a non-zero demand variance, which is used to reflect the intuitive degree of variability in demand. The larger the ratio, the more drastic the variability.

[0060] Relative variability is the ratio of the square of the non-zero demand standard deviation to the square of the non-zero demand mean, which is used to reflect the relative degree of demand variability (i.e., considering both the non-zero demand standard deviation and the non-zero demand mean). The larger the ratio, the more drastic the variability.

[0061] (4) Based on the intermittent nature, absolute variability, and relative variability, the demand type can be further calculated and a replenishment strategy can be recommended.

[0062] First, based on the above indicators, we exclude some extreme cases. When the number of non-zero demand periods is less than or equal to the limit value, it means that the demand is very sparse, and the demand type can be determined as "extremely sparse". Therefore, the safety stock of the facility-product combination is not calculated, and the recommended replenishment strategy is "make-to-order"; when the non-zero demand mean is lower than the limit value, it means that the demand is sluggish, and the demand type can be determined as "extremely sluggish", and the safety stock is not calculated, and the replenishment strategy is not recommended; when the relative variability is greater than or equal to the limit value, it means that the demand fluctuates greatly, and the safety stock may tend to be infinite, which is meaningless in reality. Therefore, the demand type can be determined as "extremely volatile", and the safety stock is not calculated, and the replenishment strategy is not recommended. After excluding the above extreme cases, the following begins to determine the demand type and recommend replenishment strategies within the normal range.

[0063] First, define the critical value of intermittency (the determination of the critical value is mainly based on the rule of thumb). If it is greater than or equal to the critical value, it is considered "discontinuous", and if it is less than the critical value, it is considered "continuous"; then define the critical values ​​of absolute variability and relative variability based on intermittency. When the absolute variability is greater than or equal to the critical value, it is judged as "fluctuation", and if it is less than the critical value, it is judged as "stable"; and when the relative variability is greater than or equal to the critical value, it is considered "violent", and if it is less than the critical value, it is considered "gentle". The critical values ​​of absolute variability and relative variability under "discontinuous" and "continuous" intermittency can be set to the same or different values.

[0064] After the intermittent nature, absolute variability, and relative variability are determined, the demand type analysis and replenishment strategy recommendations are conducted respectively.

[0065] (6) When and only when the intermittent nature is “continuous” and the relative variability is “mild”, and the demand type is accordingly classified as “smooth”, the demand distribution can be regarded as a normal distribution. At this time, the error of the classic safety stock calculation method is small, which is also easy to understand intuitively.

[0066] When demand is intermittent and changes dramatically in time series, if the safety stock calculation formula corresponding to the normal distribution is still used, the description of relative variability is particularly lacking, and the safety stock calculation is therefore inaccurate. The gamma distribution has shape parameters and scale parameters, so the gamma distribution is more suitable for describing various distributions with intermittent "discontinuity" and variability ranging from "mild" to "intense".

[0067] Specifically, when the intermittency is "continuous" and the relative variability is "intense", the demand type can be determined as "unstable", and the demand is suitable for fitting with the gamma distribution. When the intermittency is "discontinuous" and the relative variability is "gentle", the demand type can be determined as "slow", and the demand is suitable for fitting with the gamma distribution. When the intermittency is "discontinuous" and the relative variability is "intense", the demand type can be determined as "blocky", and both the gamma distribution and the negative binomial distribution are applicable.

[0068] Step 2: Recommend replenishment strategies based on demand type.

[0069] When the demand type is "smooth", the recommended replenishment strategy is the "R, Q" strategy. When the demand type is "unstable", the recommended replenishment strategy is the "s, S" strategy. When the demand type is "slow" and the absolute variability is "stable", the recommended replenishment strategy is the "R, Q" strategy (if the replenishment batch is limited to 1 product, then the "Basestock" strategy is recommended). When the demand type is "blocked", the recommended replenishment strategy is the "R, Q" strategy (if the replenishment batch is limited to 1 product, then the "T, S" strategy is recommended). When the demand type is "slow" and the absolute variability is "fluctuating", the recommended replenishment strategy is the "s, S" strategy.

[0070] Common replenishment strategies include the “R, Q” strategy, the “s, S” strategy, the “Basestock” strategy, and the “T, S” strategy.

[0071] The "R, Q" strategy is a common inventory strategy based on fixed capacity and feasible total inventory level. In the "R, Q" strategy, "R" stands for the reorder point, which is the inventory level that triggers an order when the inventory level drops to a certain level; "Q" stands for the order quantity, which is the fixed quantity for each order. When the inventory level drops below R, the system will order according to the set Q value to replenish the inventory.

[0072] The "s, S" strategy is a commonly used inventory control strategy. The "s, S" strategy means that at the beginning of each cycle, the system will check the current inventory level. If the inventory level is lower than or equal to the reorder point s, the system will place an order to increase the inventory level to the maximum inventory level S; if the inventory level is higher than s, no order will be placed. Among them, s is the reorder point, which is a constant and represents the lower limit of the inventory that triggers the order; S is the maximum inventory level, which is also a constant and represents the upper limit of the inventory that is expected to be reached after the order is placed.

[0073] The "Basestock" strategy is an inventory management method that is mainly used to determine the base or minimum inventory level that an enterprise should maintain under uncertain demand. The core idea of ​​the "Basestock" strategy is to minimize inventory holding costs by setting a minimum inventory base while ensuring service levels. This inventory base is designed to meet future expected demand and buffer uncertainty. Compared with the "s, S" strategy, the main difference of the "Basestock" strategy is whether fixed ordering costs are considered. The "Basestock" strategy is optimal when the fixed ordering cost is zero and the order quantity is infinite. When there are fixed ordering costs, the "s, S" strategy may be more appropriate.

[0074] The "T, S" strategy is a commonly used inventory management method. The "T, S" strategy means that the inventory is checked every fixed period T, and an order is issued to restore the inventory level to the maximum inventory level S. If the inventory level at the time of the check is I, the order quantity is SI. This strategy does not set an order point, but only a fixed inspection period and a maximum inventory level.

[0075] Step three, perform safety stock calculation, and use the calculation method proposed in the present invention for non-normal demand distribution.

[0076] After completing the fitting of the demand distribution, except for the demand classified as normal distribution, the safety stock calculation can be performed according to the mature formula. When the demand distribution is classified as gamma distribution or negative binomial distribution, there is no formula to refer to. Therefore, the present invention proposes a safety stock calculation method for non-normal demand distributions such as gamma distribution and negative binomial distribution.

[0077] First, we review the safety stock calculation formula under normal distribution and analyze its characteristics. The existing safety stock calculation mainly assumes that customer demand is randomly distributed and conforms to the normal distribution, and then uses the classic safety stock formula, that is, Where SS is the safety stock, is the average lead time, is the daily average demand, z is the number of standard deviations under a certain service level, σ d is the standard deviation of daily demand d, σL is the standard deviation of the lead time L. When the accuracy of safety stock calculation is not high, it is the standard deviation of the lead time L L is often set to 0, so the formula is simplified to However, using normal distribution to fit customer demand is a simplistic operation that lacks the ability to characterize the relative variability of demand. In fact, normal distribution is only applicable to situations where the distribution is continuous and the relative variability is small. It is not applicable to situations where there are widespread demand gaps or where further characterization of demand variability is required.

[0078] In actual operation, in order to more accurately characterize the characteristics of the distribution, the daily average demand of non-zero demand is used and standard deviation The normal distribution can be written as Norm(μ,σ), where Safety stock SS reflects the inventory level that needs to be achieved to meet the service level of demand. The z in the formula needs to be solved by the percent point function (PPF).

[0079] When the demand distribution is a gamma distribution, it can be recorded as Gamma (α, β), where α is the shape parameter, which mainly determines the shape of the distribution curve, especially the position of the peak and the sharpness of the curve; β is called the scale parameter, which mainly determines the width of the distribution, and its reciprocal is the inverse scale parameter θ, that is, According to the characteristics of the gamma distribution, its mean μ = α * β, variance V = α * β 2 . And the mean is known Standard Deviation Variance It can be further deduced that Then we need to further derive the safety stock calculation method based on the quantile function of the gamma distribution, with the inputs being α, α, and the service level service level , various existing programming languages ​​can be used to first calculate PPF (α, β, service level ), and then use actual data observation or simulation methods to make corrections and adjustments to obtain an accurate and applicable safety stock calculation method.

[0080] When the demand distribution is a negative binomial distribution, it can be recorded as Nbinom(r,p). Negative binomial distribution represents a series of independent experiments, each of which has two results: success and failure. The probability of success is constant, recorded as p, and the experiment continues until it succeeds r times, and the number of successes is recorded as r. According to the characteristics of negative binomial distribution, its mean is usually recorded as or Considering that the number of successes is usually a large positive integer, the mean is further simplified to The variance is usually or We use Known mean Standard Deviation Variance Further deducing Then we need to further derive the safety stock calculation method based on the quantile function of the negative binomial distribution, with the inputs being r, p, and the service level service level , you can use various existing programming languages ​​to first find the PPF (r, p, service level ), and then use actual data observation or simulation methods to make corrections and adjustments to obtain an accurate and applicable safety stock calculation method.

[0081] The following are examples of specific applications for illustration, in which all parameters are empirical values. According to the demand time series generation process, the following Table 1 shows the processed demand time series.

[0082] Table 1 Demand time series table

[0083]

[0084] The input table is the demand time series of a certain product of a certain customer. The demand values ​​have been summarized by unit time interval and arranged in order of occurrence. The larger the period number, the later the demand occurs. After being transmitted from the customer-product combination to the facility-product combination, the facility-product demand time series is also like this table.

[0085] Non-zero demand parameter calculation is based on the empirical formula mentioned above:

[0086] Non-zero demand period nz =9, non-zero demand mean Non-zero demand standard deviation Intermittent calculation: Variability Calculation: Absolute Variability Relative variability

[0087] Extreme Case Checks:

[0088] 1) When the non-zero demand period period nz ≤3, the demand is judged to be "extremely sparse";

[0089] 2) When the non-zero demand mean Demand was judged to be "extremely depressed";

[0090] 3) When relative variability re_variability>5, the demand is judged as "extremely volatile".

[0091] The current case does not fall into any of the above extreme situations.

[0092] Intermittency setting threshold: The intermittency threshold can be set to 1.32. When the intermittency is greater than or equal to 1.32, the demand is determined to be "discontinuous". When the intermittency is less than 1.32, the demand is determined to be "continuous". The intermittency of the current case is 2.11, so the demand is "discontinuous".

[0093] Absolute variability threshold: The absolute variability threshold can be uniformly set to 10. ab When it is greater than or equal to 10, the demand is judged to be "fluctuating", absolute variability ab When it is less than 10, the demand is considered to be "stable". ab It is 67.12, so the demand is "fluctuating".

[0094] Relative variability threshold: The relative variability threshold can be uniformly set to 0.49. re When it is greater than or equal to 0.49, the demand is judged to be "intense", and the relative variability re When it is less than 0.49, the demand is judged to be "relaxed". re It is 1.46, so the demand is “intense”.

[0095] Demand type determination: The intermittent of the current case is "discontinuous", and the relative variability is "variability". re is "intense", so the demand type class d To "block".

[0096] Replenishment strategy recommendation: The current demand type is "block", and the recommended replenishment strategy is "R, Q" strategy (if the replenishment batch is limited to 1 product, then the recommended strategy is "T, S" strategy)

[0097] Distribution fitting: Since the current case requirement type is "block", gamma distribution or negative binomial distribution fitting can be used.

[0098] Importing lead time information: Lead time mean Lead time standard deviation σ L =0.

[0099] Lead time demand variable preparation: lead time demand mean The standard deviation of demand in advance,

[0100] Service level setting: Set the service level service_level = 0.95.

[0101] Gamma distribution parameter estimation: Gamma(α,β), where the shape parameter Scale parameter

[0102] Negative binomial distribution parameter estimation: Nbinom(r,p), where the number of successes Probability of success

[0103] Percent Point Function (PPF) calculation: Use programming tools such as Python and Matlab to call the corresponding PPF library. For the gamma distribution, the input parameters are service level sev_level, shape parameter α, and scale parameter β. For the negative binomial distribution, the input parameters are service level sev_level, number of successes r, and probability of success p.

[0104] Step 4: Adjust and calibrate the safety stock formula through simulation. After obtaining the quantile function value, it is necessary to further adjust the calculation method of the safety stock quantity according to the actual data observation value or simulation, and the shortage and overstocking situation in the specific scenario.

[0105] The simulation can be performed using professional simulation software, such as Matlab, to set the demand time series, replenishment strategy, and safety stock. The replenishment strategy and safety stock are calculated by the replenishment recommendation strategy device and safety stock calculation device of the patent through the demand time series to obtain the recommended replenishment strategy and safety stock value. After the simulation is completed, the frequency and severity of out-of-stock and overstock are recorded. If the product has valuable information, the out-of-stock cost and overstock cost can also be calculated. If the safety stock calculation device determines that the demand distribution of a facility-product combination is non-normal, if there is a high probability of severe out-of-stock or overstock, data analysis is required to correct the safety stock calculation method.

[0106] Compared with the prior art, the beneficial effect of the present invention is that according to the technical solution, by calculating intermittency, absolute variability, relative variability and demand type during demand analysis, customer demand can be fitted into normal distribution, gamma distribution and negative binomial distribution, and based on the idea that gamma distribution and negative binomial distribution can characterize relative variability, the safety stock calculation method of gamma distribution and negative binomial distribution is fitted. This solution is combined with the dynamic programming method to calculate the safety stock optimization problem. In practical applications, it can better calculate the safety stock that each facility-product node should reserve, reduce out-of-stock phenomena, avoid inventory waste, and thus reduce inventory costs. This solution automatically recommends replenishment strategies, which can help each node in the supply chain better carry out replenishment and procurement plans.

Claims

1. A safety stock allocation and replenishment strategy recommendation method, characterized in that: include: Design time series characteristic parameter step, based on the analysis of demand characteristics, design parameters to describe the demand time series characteristics, Determine the demand distribution type step, determine the demand distribution type based on intermittency, absolute variability, and relative variability; Recommend replenishment strategy steps, and recommend replenishment strategies based on the demand type of demand distribution; The safety stock calculation steps are based on the results of demand distribution judgment as the initial judgment basis for safety stock calculation. Combining statistics and actual data, the safety stock calculation method under non-normal distribution is derived.

2. A safety stock allocation and replenishment strategy recommendation method as claimed in claim 1, characterized in that: The parameters designed to describe the characteristics of the demand time series include intermittency, absolute variability, relative variability, and demand type.

3. A safety stock allocation and replenishment strategy recommendation method as claimed in claim 1, characterized in that: The step of determining the demand distribution type embeds a built-in algorithm, namely, a method for calculating intermittency, absolute variability, relative variability, and demand type.

4. A safety stock allocation and replenishment strategy recommendation method as claimed in claim 1, characterized in that: The step of designing time series characteristic parameters excludes extreme demand time series that are not worthy of calculating safety stocks or recommending replenishment strategies by setting limit values ​​for the number of non-zero demand periods and the non-zero demand mean.

5. A safety stock allocation and replenishment strategy recommendation method as claimed in claim 1, characterized in that: The determination demand distribution type is normal distribution, gamma distribution or negative binomial distribution.

6. A safety inventory configuration and replenishment strategy recommendation method as claimed in claim 1, characterized in that: The safety stock calculation step embeds the safety stock calculation method under normal distribution, gamma distribution and negative binomial distribution, that is, the quantile function PPF is calculated as the basic value of the safety stock.

7. A method for recommending safety inventory configuration and replenishment strategy according to claim 6, characterized in that: The calculation method of safety stock under the gamma distribution and negative binomial distribution is: For the gamma distribution, calculate the shape parameter α and scale parameter β and set the service level service level , according to the quantile function PPF, a basic value of safety stock is obtained, that is, PPF (α, β, service level ); For the negative binomial distribution, calculate the success probability p and the number of successes r and set the service level service level , according to the quantile function PPF, a basic value of safety stock is obtained, that is, PPF (r, p, service level ).

8. A safety inventory configuration and replenishment strategy recommendation method as claimed in claim 1, characterized in that: It also includes adjustment and correction steps, using professional simulation software to set the demand time series, replenishment strategy, and safety stock, wherein the replenishment strategy and safety stock, that is, the recommended replenishment strategy and safety stock value calculated through the demand time series through the replenishment recommendation strategy step and the safety stock calculation step, are used to test the effectiveness of the safety stock setting.

9. A method for recommending safety inventory configuration and replenishment strategy according to claim 8, characterized in that: Also includes, After the simulation software is finished running, the frequency and severity of out-of-stock and overstocking are recorded. If the product has valuable information, the out-of-stock cost and overstocking cost can also be calculated. If the safety stock calculation device determines that the demand distribution of a facility-product node is non-normal and there is a high probability of serious out-of-stock or overstocking, data analysis is performed to correct the safety stock calculation method.

10. A device using the safety inventory configuration and replenishment strategy recommendation method according to any one of claims 1 to 9, characterized in that: include: The demand distribution determination module designs parameters such as intermittency, absolute variability, relative variability, and demand type based on the analysis of demand characteristics; Demand distribution type determination module, which determines the demand distribution type based on intermittency, absolute variability, and relative variability; The replenishment strategy recommendation module recommends replenishment strategies based on demand distribution type; The safety stock calculation module uses the result of demand distribution as the initial judgment basis for safety stock calculation, and combines statistics and actual data to derive the safety stock calculation method under non-normal distribution; The safety stock correction module uses simulation methods to determine the applicability and adjustment direction of the safety stock calculation method based on out-of-stock and over-stock situations.

Images