WO2023082315A1 - 电子固废回收全流程智能解析方法及系统 - Google Patents

电子固废回收全流程智能解析方法及系统 Download PDF

Info

Publication number
WO2023082315A1
WO2023082315A1 PCT/CN2021/131990 CN2021131990W WO2023082315A1 WO 2023082315 A1 WO2023082315 A1 WO 2023082315A1 CN 2021131990 W CN2021131990 W CN 2021131990W WO 2023082315 A1 WO2023082315 A1 WO 2023082315A1
Authority
WO
WIPO (PCT)
Prior art keywords
recycling
model
solid waste
electronic solid
multivariate
Prior art date
Application number
PCT/CN2021/131990
Other languages
English (en)
French (fr)
Inventor
栾小丽
成程
孙晓安
刘飞
Original Assignee
江南大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 江南大学 filed Critical 江南大学
Publication of WO2023082315A1 publication Critical patent/WO2023082315A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00Administration; Management
    • G06Q10/06Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/067Enterprise or organisation modelling
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00Administration; Management
    • G06Q10/04Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00Administration; Management
    • G06Q10/30Administration of product recycling or disposal
    • YGENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02WCLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO WASTEWATER TREATMENT OR WASTE MANAGEMENT
    • Y02W90/00Enabling technologies or technologies with a potential or indirect contribution to greenhouse gas [GHG] emissions mitigation

Definitions

  • the invention relates to the field of intelligent information interaction technology, in particular to an intelligent analysis method and system for the whole process of electronic solid waste recycling.
  • the current electronic solid waste recycling has entered the Internet + recycling mode.
  • the process is that consumers place orders online, and recyclers come to collect them at recycling outlets, and the recycling outlets send them to sorting centers for sorting, and finally send them to external dismantling companies for demolition.
  • the information interaction between the various links of the existing electronic solid waste recycling process is not smooth or even digitized, and the existing data has not been analyzed in depth, which further leads to low efficiency of the recycling process, and even the existing analysis is basically only for the recycling process.
  • the correlation of data between various processes is ignored, and the value of historical data is not fully utilized, resulting in a mediocre analysis effect and failure to improve recycling efficiency.
  • the technical problem to be solved by the present invention is to overcome the defects existing in the prior art, provide an intelligent analysis method and system for the whole process of electronic solid waste recycling, and improve the degree of information interaction among the participants in the whole process of electronic solid waste recycling, It reduces the recycling cost, realizes the informatization and intelligence of electronic solid waste recycling, improves the efficiency and scientificity of electronic solid waste recycling network construction, saves the recycling time of offline recyclers, and reduces the freight in the process of vehicle transportation. Reduce the cost of supply and transfer between the sorting center and the dismantling enterprise, thereby increasing the amount of recycling and satisfaction.
  • the present invention provides an intelligent analysis method for the whole process of electronic solid waste recycling, including:
  • S1 Obtain the historical data of electronic solid waste generation in the area to be predicted and the historical data affecting the solid waste generation, perform normalization processing on the historical data, use the normalized historical data to construct a multivariate gray model, and use intelligent learning
  • the algorithm performs error compensation on the predicted value of the multivariate gray model to obtain the predicted value of electronic solid waste generation in the area to be predicted;
  • S2 Construct a site selection model for recycling outlets based on the predicted value of electronic solid waste generation, use intelligent algorithms to solve the location selection model for recycling outlets, obtain the location distribution of recycling outlets, and build a sorting center location model based on the location distribution of recycling outlets , use the intelligent algorithm to solve the sorting center location model, and obtain the location distribution of the sorting center;
  • S3 Construct a recycling personnel scheduling model based on the location distribution of the recycling outlets, construct a recycling vehicle scheduling model based on the location distribution of the recycling outlets and the sorting center, construct a long-term supply model based on the location distribution of the sorting center, and
  • the short-term cargo transfer model uses intelligent algorithms to solve the recycling personnel scheduling model, recycling vehicle scheduling model, long-term supply model and short-term cargo scheduling model, and obtains the recycling personnel scheduling scheme, recycling vehicle scheduling scheme, long-term supply scheme and short-term scheduling model.
  • cargo plan
  • S4 Perform visual processing on the recycling personnel scheduling plan, recycling vehicle scheduling plan, long-term supply plan and short-term cargo transfer plan.
  • the normalized historical data is used to construct a multivariate gray model
  • the intelligent learning algorithm is used to compensate the error of the predicted value of the multivariate gray model, so as to obtain the predicted value of electronic solid waste generation in the area to be predicted
  • represents the delay time
  • r represents the historical sequence number
  • t represents the sequence number
  • a cumulative result of is the cumulative result of Y G (0)
  • Y G (0) is the output prediction sequence of the multivariate gray model
  • both a and u are model parameters
  • a represents the control coefficient
  • u represents the gray action
  • rt represents the time series number to be predicted
  • ⁇ (0) Y (0) - Y G (0)
  • Y G (0) ⁇ Y G (0) ( ⁇ +1),Y G (0) ( ⁇ +2),...,Y G (0) ( ⁇ +r) ⁇ ,
  • ⁇ (0) ⁇ (0) ( ⁇ +1), ⁇ (0) ( ⁇ +2),..., ⁇ (0) ( ⁇ +r) ⁇ ;
  • S15 Use the intelligent learning method to intelligently compensate the prediction error of the multivariate gray model, and obtain the error compensation sequence
  • Y H (0) represents the final intelligent prediction result of electronic solid waste generation
  • the method for constructing a site selection model for recycling outlets based on the predicted value of electronic solid waste generation includes:
  • a site selection model for recycling sites was established with the goal of minimizing economic costs and maximizing the coverage of solid waste generation centralized points.
  • the method for constructing the location model of the sorting center according to the location distribution of the recycling outlets includes:
  • the method for constructing a recycling personnel scheduling model based on the location distribution of the recycling outlets includes:
  • the location distribution of the recovery network is used as an input, and the recovery personnel scheduling scheme is used as an output to construct a recovery personnel scheduling model.
  • constructing a recycling vehicle scheduling model based on the location distribution of the recycling outlets and the location distribution of the sorting center includes:
  • the location distribution of recycling outlets after normalization processing and the location distribution of the sorting center are used as input, and the recycling vehicle scheduling scheme is used as output to construct a recycling vehicle scheduling model.
  • the present invention also provides an intelligent analysis system for the whole process of electronic solid waste recycling, including:
  • a data acquisition module the data acquisition module is used to acquire the amount of electronic solid waste in the area to be predicted and the historical data affecting the amount of solid waste, perform normalization processing on the historical data, and use the historical data after normalization processing
  • Data constructs a multivariate gray model uses intelligent learning algorithms to compensate for errors in the predicted values of the multivariate gray model, and obtains the predicted value of electronic solid waste generation in the area to be predicted;
  • a location distribution planning module the location distribution planning module is used to construct a site selection model for recycling outlets based on the predicted value of electronic solid waste generation, and use an intelligent algorithm to solve the location selection model for recycling outlets to obtain the location distribution of recycling outlets, and according to the The location distribution of recycling outlets builds a sorting center location model, uses intelligent algorithms to solve the sorting center location model, and obtains the location distribution of the sorting center;
  • An intelligent scheduling module is used to construct a recycling personnel scheduling model based on the location distribution of the recycling outlets, construct a recycling vehicle scheduling model based on the location distribution of the recycling outlets and the location distribution of the sorting center, and construct a recycling vehicle scheduling model based on the location distribution of the sorting centers.
  • Long-term supply model and short-term cargo transfer model are built based on the distribution of central locations, and intelligent algorithms are used to solve the recycling personnel scheduling model, recycling vehicle scheduling model, long-term supply model and short-term cargo scheduling model, and the recycling personnel scheduling plan and recycling vehicle scheduling are obtained.
  • plan long-term supply plan and short-term transfer plan;
  • a visualization module the visualization module is used for visualizing the recycling personnel scheduling plan, recycling vehicle scheduling plan, long-term supply plan and short-term cargo transfer plan.
  • the data acquisition module includes an electronic solid waste prediction sub-module, and the electronic solid waste prediction sub-module is used to construct a multivariate gray model using normalized historical data, and uses an intelligent learning algorithm to Error compensation is performed on the predicted value of the multivariate gray model to obtain the predicted value of electronic solid waste generation in the area to be predicted, including:
  • represents the delay time
  • r represents the historical sequence number
  • t represents the sequence number
  • a cumulative result of is the cumulative result of Y G (0)
  • Y G (0) is the output prediction sequence of the multivariate gray model
  • both a and u are model parameters
  • a represents the control coefficient
  • u represents the gray action
  • rt represents the time series number to be predicted
  • ⁇ (0) Y (0) - Y G (0)
  • Y G (0) ⁇ Y G (0) ( ⁇ +1),Y G (0) ( ⁇ +2),...,Y G (0) ( ⁇ +r) ⁇ ,
  • ⁇ (0) ⁇ (0) ( ⁇ +1), ⁇ (0) ( ⁇ +2),..., ⁇ (0) ( ⁇ +r) ⁇ ;
  • S15 Use the intelligent learning method to intelligently compensate the prediction error of the multivariate gray model, and obtain the error compensation sequence
  • Y H (0) represents the final intelligent prediction result of electronic solid waste generation
  • the location distribution planning module includes a recycling network construction sub-module, and the recycling network construction sub-module is used to construct a recycling network site selection model based on the predicted value of electronic solid waste generation, including:
  • a site selection model for recycling sites was established with the goal of minimizing economic costs and maximizing the coverage of solid waste generation centralized points.
  • the location distribution planning module includes a sorting center construction sub-module, and the sorting center construction sub-module is used to construct a sorting center site selection model according to the location distribution of the recycling network, including:
  • the present invention intelligently analyzes the whole process of electronic solid waste recycling, fully taps the value in historical data, builds a model in a targeted manner, uses an intelligent algorithm to solve the model, and improves the information of each participant in the whole process of electronic solid waste recycling
  • the degree of interaction reduces the cost of recycling, realizes the informatization and intelligence of electronic solid waste recycling, improves the efficiency and scientificity of electronic solid waste recycling network construction, saves the recycling time of offline recyclers, and reduces the time spent on vehicle transportation.
  • the freight costs are reduced, and the cost of supply and transfer between the sorting center and the dismantling enterprise is reduced, thereby increasing the amount of recycling and satisfaction.
  • Fig. 1 is a schematic flowchart of a whole-process intelligent analysis method for electronic solid waste recycling according to an embodiment of the present invention.
  • Fig. 2 is a schematic diagram of the hardware structure of a whole-process intelligent analysis system for electronic solid waste recycling according to an embodiment of the present invention.
  • Fig. 3 is the embodiment of the present invention
  • this embodiment provides an intelligent analysis method for the whole process of electronic solid waste recycling, including:
  • S1 Obtain the historical data of electronic solid waste generation in the area to be predicted and the historical data affecting the solid waste generation, perform normalization processing on the historical data, use the normalized historical data to construct a multivariate gray model, and use intelligent learning
  • the algorithm performs error compensation on the predicted value of the multivariate gray model to obtain the predicted value of electronic solid waste generation in the area to be predicted;
  • S2 Construct a site selection model for recycling outlets based on the predicted value of electronic solid waste generation, use intelligent algorithms to solve the location selection model for recycling outlets, obtain the location distribution of recycling outlets, and build a sorting center location model based on the location distribution of recycling outlets , use the intelligent algorithm to solve the sorting center location model, and obtain the location distribution of the sorting center;
  • S3 Construct a recycling personnel scheduling model based on the location distribution of the recycling outlets, construct a recycling vehicle scheduling model based on the location distribution of the recycling outlets and the sorting center, construct a long-term supply model based on the location distribution of the sorting center, and
  • the short-term cargo transfer model uses intelligent algorithms to solve the recycling personnel scheduling model, recycling vehicle scheduling model, long-term supply model and short-term cargo scheduling model, and obtains the recycling personnel scheduling scheme, recycling vehicle scheduling scheme, long-term supply scheme and short-term scheduling model.
  • cargo plan
  • S4 Visualize the recovery personnel scheduling scheme, recovery vehicle scheduling scheme, long-term supply scheme and short-term cargo transfer scheme.
  • step S1 the normalized historical data is used to construct a multivariate gray model, and the intelligent learning algorithm is used to make errors in the predicted values of the multivariate gray model. Compensation, the methods to obtain the predicted value of electronic solid waste generation in the area to be predicted include:
  • S11 Use the normalized historical data to construct a multivariate gray model as follows (1). Specifically, construct a multivariate gray model based on the historical data of electronic solid waste generation, population density, and household disposable income.
  • Y (0) ⁇ Y (0) ( ⁇ +1), Y (0) ( ⁇ +2), ..., Y (0) ( ⁇ +r) ⁇ ;
  • represents the delay time
  • r represents the historical sequence number
  • t represents the sequence number
  • a cumulative result of is the cumulative result of Y G (0)
  • Y G (0) is the output prediction sequence of the multivariate gray model
  • both a and u are model parameters
  • a represents the control coefficient
  • u represents the gray action
  • rf represents the time series number to be predicted
  • ⁇ (0) Y (0) - Y G (0) (5)
  • Y G (0) ⁇ Y G (0) ( ⁇ +1),Y G (0) ( ⁇ +2),...,Y G (0) ( ⁇ +r) ⁇ ,
  • ⁇ (0) ⁇ (0) ( ⁇ +1), ⁇ (0) ( ⁇ +2),..., ⁇ (0) ( ⁇ +r) ⁇ ;
  • S15 Use the intelligent learning method to intelligently compensate the prediction error of the multivariate gray model, and obtain the error compensation sequence
  • Y H (0) represents the final intelligent prediction result of electronic solid waste generation
  • the intelligent learning method may be: neural network, support vector machine, expert system, fuzzy system, deep learning, reinforcement learning.
  • step S2 the method of constructing the site selection model of recycling outlets based on the predicted value of electronic solid waste generation includes:
  • a site selection model for recycling sites was established with the goal of minimizing economic costs and maximizing the coverage of solid waste generation centralized points.
  • step S2 after constructing the site selection model of recycling sites, the intelligent algorithm is used to solve the site selection model of recycling sites to obtain the location distribution of recycling sites.
  • the coefficient of is obtained to obtain information such as the number of personnel and vehicles configured by the recycling outlets.
  • step S2 the predicted value of electronic solid waste generation is obtained from the prediction in step S1, and the economic index data include: construction cost, operating cost, transportation cost and cargo damage cost.
  • step S2 the method for constructing a sorting center location model according to the location distribution of the recycling outlets includes:
  • step S2 after constructing the sorting center location model, the intelligent algorithm is used to solve the sorting center location model to obtain the location distribution of the sorting center, and according to the recycling amount of the recycling outlets covered by the sorting center, according to a certain
  • the coefficient of is obtained to obtain information such as the number of personnel and vehicles configured by the recycling outlets.
  • the upper layer mainly considers: economic indicators and solid waste generation forecast results, etc.
  • the lower layer mainly considers the satisfaction and economic indicators of the dismantling enterprise.
  • the upper-level economic indicators include construction costs, operating costs, transportation costs, and cargo damage costs
  • the lower-level economic indicators include transaction prices, transportation costs, and cargo damage costs.
  • the upper layer mainly considers the cost control of the decision-making layer, that is, from the residents’ recycling to the upper layer of the sorting center, the main costs considered include construction costs, operating costs, and variable investment costs, of which variable investment costs mainly Including transportation costs, cargo damage costs, etc. Therefore, the expression of the upper model can be obtained as follows:
  • constraint (10) means that the sum of investment costs of all sorting centers has an upper limit
  • constraint (11) means that at least one alternative sorting center is selected
  • constraint (12) means that the inventory of the sorting center must be less than The maximum inventory
  • constraint (13) guarantees that the solution of the upper model is 0-1 type.
  • the lower layer mainly considers that the backend of the present invention, that is, the dismantling enterprise, is a third-party enterprise and cannot control it. Therefore, we mainly consider for the dismantling enterprises to achieve the highest service satisfaction for dismantling enterprises under the objective function of ensuring the lowest cost. Therefore, the main things that need to be considered are price, distance and damage to the goods. So the expression of the lower model is as follows:
  • the constraint formula (16) ensures that the solution obtained by the lower model is a 0-1 value
  • the constraint formula (17) indicates that the sorting center cooperates with the dismantling enterprise to establish a sorting center in the area
  • the constraint formula (18) expresses The mode of cooperation between the sorting center and the dismantling enterprise, 1 represents cooperation, otherwise it is 0, and the constraint (19) indicates that the distance between the sorting center and the dismantling enterprise cannot exceed the maximum distance.
  • step S2 for the solution of the model, especially for the small-scale model, the number of variables is small, and the upper model can list all the decision-making schemes through enumeration methods such as the branch and bound method, and use them as the lower-level decision-making
  • the solution of the variable is input into the lower layer model, and then the MATLAB function toolbox is used to solve it, and then the calculated (Q, B) is returned to the upper layer objective function, the cost of the corresponding decision-making plan is calculated, and the lowest-cost location plan is screened out.
  • step S3 the method of constructing a recycling personnel scheduling model based on the location distribution of the recycling network includes:
  • the location distribution of the recovery network is used as an input, and the recovery personnel scheduling scheme is used as an output to construct a recovery personnel scheduling model.
  • step S3 the construction of the recovery personnel scheduling model is also related to the order information, for example, as follows:
  • the status information of offline recyclers screens out the recyclers who can serve; among the screened recyclers, they are comprehensively sorted according to the current indicators such as small number, proximity to customers, and high proficiency; use the sorting information to increase recycling for the first-ranked recyclers Task, update the service time list of recyclers; when the user deletes the order or the recycler fails to pick up the package, set the time list of the relevant recyclers to zero, and update the service time list of the recyclers; construct the service time list of the recyclers as Input, the recovery personnel scheduling model with the recovery personnel scheduling scheme as the output.
  • the online order information includes the type of goods, whether there is an elevator, the volume of the goods, the weight of the goods, the recycling location, whether to divide into batches, special requirements, etc.;
  • the status information of the offline recycling personnel includes the service time list of the recycling personnel, Recyclers' current location, weather changes, traffic conditions, road conditions, vehicle performance, etc.
  • the goal of the vehicle routing optimization problem is to reasonably arrange the transportation vehicles and determine the transportation route of each vehicle, so as to ensure the shortest transportation distance. Additionally, given the following constraints:
  • Each customer is delivered by only one vehicle:
  • y ik means that the k-th vehicle is responsible for the transportation of the i-th customer, i means the i-th customer, i ⁇ V; k means the k-th car, k ⁇ K.
  • Vehicle k travels from customer i to customer j:
  • x ijk means that vehicle k travels from customer i to customer j.
  • the total weight of goods transported on each route shall not exceed the maximum load capacity of the vehicle:
  • g i represents the i-th customer's stock of goods, Indicates the total delivery volume for each route.
  • d ij represents the distance from customer i to customer j.
  • Step 1 On the premise that the geographical location of each city is known, calculate the shortest distance from the recycling outlet to the customer and between different customers.
  • the solution to the vehicle routing optimization problem is a set of routes that traverse all customers. Let the coordinates of two customers be (x i , y i ), (x j , y j ), where x i represents the abscissa of customer i’s position, y i represents the ordinate, and calculate the distance c ij :
  • Step 2 Calculate the mileage saved, which specifically includes the following steps:
  • Step 2.1 Calculate the saved mileage P ij :
  • P ij represents the mileage saved by customer i and customer j
  • c ij represents the distance from customer i to customer j
  • c i,0 represents the straight-line distance between customer i and the recycling outlet
  • c 0,j represents the distance between customer j and customer j.
  • One of s i and s j is the starting point of the path, and the other is the end point of the path;
  • step 3 If one of the above conditions is met, go to step 3, otherwise, delete the current P ij from S.
  • Step 4 Repeat steps 2 and 3 until S is an empty set.
  • step S3 constructing a recycling vehicle scheduling model based on the location distribution of the recycling outlets and the location distribution of the sorting center includes:
  • the location distribution of recycling outlets after normalization processing and the location distribution of the sorting center are used as input, and the recycling vehicle scheduling scheme is used as output to construct a recycling vehicle scheduling model.
  • the vehicle information of the sorting center includes: fuel consumption, carrying capacity, maximum speed, vehicle type and maximum driving distance, etc.
  • constraints that the model needs to satisfy include:
  • g i represents the inventory of the i-th customer's goods
  • vi represents the volume of the i-th customer's goods
  • x ijk means that vehicle k travels from customer i to customer j.
  • y ik means that the k-th vehicle is responsible for the transportation of the i-th customer, i means the i-th customer, i ⁇ V; k means the k-th car, k ⁇ K.
  • the objective function is:
  • Goal 2 The minimum number of delivery vehicles dispatched at one time:
  • Step 1 Randomly generate the population (x 1 , x 2 , ..., x n ), and judge whether the individuals in the current population meet the requirements of the maximum load capacity and the minimum vehicle;
  • Step 2 Calculate the objective function values f 1 ( xi ) and f 2 ( xi ) and generate a weight vector;
  • Step 3 Calculate the Euclidean distance between any two weight vectors, and find T nearest weight vectors.
  • Step 4 randomly select two solutions from B(i), perform a crossover operation to generate a child y, and then perform a mutation operation on y to generate y';
  • Step 5 judge whether y' meets the weight limit requirement, if not, delete y' and return to step 4;
  • Step 6 Store the offspring and update the external population (abbreviated EP);
  • Step 7 Judging whether the termination condition is reached, if not, return to step 5.
  • step S3 the method for constructing a long-term supply model based on the location distribution of the sorting centers includes:
  • the evaluation data includes the type of goods, the scope of dismantling, the receiving price, the transaction reputation of the sorting center, the brand range, whether self-owned vehicles, supply price, delivery speed, and the transaction reputation of the dismantling enterprise, etc. .
  • the sorting center can only be matched with one dismantling center, and the dismantling center theoretically has no limit on the number of sorting centers, that is, one-to-many bilateral matching.
  • mapping ⁇ in long-term supply: U ⁇ V ⁇ U ⁇ V, and all meet the following conditions:
  • the evaluation index system of the sorting center and dismantling center is mainly considered to give the corresponding comprehensive satisfaction value, and finally match according to the satisfaction value.
  • the evaluation system mainly involves four types of indicators, of which is a 0-1 indicator for S1 category, It is S 2 language indicators, similarly there are S 3 interval indicators and S 4 general numerical indicators.
  • the dismantling center V j For S1 type 0-1 indicators such as whether the mobile phone can be turned on, whether there are maintenance records, etc., it is necessary for the dismantling center V j to set the expected value for the pth indicator of the sorting center It is the actual value of the product in the sorting center. like Then the evaluation index value is 1, otherwise it is 0.
  • the dismantling center obtains the 0-1 evaluation matrix of the sorting center to the dismantling center according to the actual situation And the evaluation matrix of the dismantling center to the sorting center
  • the satisfaction matrix of the sorting center to the dismantling center can be obtained from formula (34) Satisfaction matrix of sorting center and dismantling center
  • S 2 language evaluation indicators such as sorting center credibility, product description compliance, etc.
  • the hard constraint means that the attribute value of one index must meet the requirements of the other party, that is, when the subject U i has a hard requirement for the index A p in the subject V, if the subject V j does not meet the hard constraint of U i , then when calculating the satisfaction Assign a very large negative value -K i , so that it cannot be matched with the subject U i ;
  • the soft constraint means that the index attribute value provided by one party can meet the expectations of the other party as much as possible, that is, the subject U i has the same effect on the index A p in the subject V The higher the satisfaction value expectation, the better. Calculated as follows:
  • formula (38) Indicates that the subject U i has a hard constraint on the index A p in the subject V, and assigns a negative value K i , and K i is large enough.
  • formula (39) Indicates that the subject U i has soft constraints on the index A p in the subject V, and K i represents the assignment under the index A p , Indicates the value of this indicator calculated according to formula 35-37.
  • each sorting center has very different needs for product recycling, it is obviously unreasonable to give a uniform weight distribution. Therefore, this paper designs a random preference order weight distribution for the waste electronic recycling process.
  • the index preference of the dismantling center to the sorting center is relatively fixed, and is generally given according to subjective experience. Specifically, as shown in Figure 3, the order of the indicators is from left to right, and the importance decreases in turn.
  • E a, E b, and E c are the indicators in the indicator system E, and the given indicator sequence 1 ⁇ g is given the weight distribution respectively. and make
  • the index ordinal value table is set according to the preference of the sorting center, and the corresponding satisfaction matrix is automatically updated and
  • the value calculated by formulas (34)-(39) is the gap between the evaluation index and the expectation. According to the distribution of weights in Figure 3, the difference matrix can be obtained as well as Its calculation formula is:
  • the final difference matrix represents the gap between the expectation and actual value of the sorting center or online store, so the transformation formula can be used to obtain the satisfaction value matrix of the sorting center for the dismantling center And the satisfaction value matrix of the dismantling center to the sorting center
  • the index information and attributes that the sorting center and the dismantling center mainly focus on are calculated according to their specific sequence value preferences, and are improved according to the calculation method of the satisfaction value proposed by the present invention, and finally the information about the classification is obtained. Satisfaction value matrix for picking centers and dismantling centers. Therefore, in order to determine the multi-objective optimization model between the sorting center and the dismantling center, a multi-objective optimization model based on the satisfaction value can be established according to the satisfaction value matrix obtained from formulas (42) and (43):
  • formulas (44) and (45) represent the objective functions that make the sorting center and the dismantling center reach the maximum satisfaction value respectively; meanwhile, the value of x ij represents the matching result of the sorting center and the dismantling center.
  • Constraint 46 means that the sorting center can only match at most one dismantling center.
  • ⁇ 1 and ⁇ 2 also show the degree of importance attached to the sorting center and dismantling center in this matching.
  • step S3 the method of constructing a short-term goods transfer model based on the location distribution of the sorting center includes:
  • step S3 it is first necessary to analyze the expectation of the supply plan.
  • the short-term supply problem of the dismantling enterprise the main cost of both the supplier and the buyer comes from the freight generated by the transportation of waste electrical and electronic products.
  • the short-term supply plan should be as short-term as possible. It is close to the demand value of the dismantling enterprise.
  • the satisfaction between the dismantling enterprise and the sorting center will also affect the feasibility of the short-term supply plan.
  • the dismantling enterprise also tends to be the same Sorting centers that have a history of cooperation cooperate again.
  • the present invention sets four goals, which are: minimizing freight, minimizing error, maximizing satisfaction of both parties and maximizing cooperation history.
  • the objective function of freight cost minimization comprehensively considers the transportation distance and the weight of the transported goods in the process of the transaction between the sorting center and the dismantling enterprise. Dismantling companies allocate more goods, and dismantling companies that are farther away allocate less goods to reduce freight costs. The larger the target value, the better the freight savings, otherwise the higher the freight.
  • the objective function to minimize the shipping cost is:
  • caf i is the freight calculated by sorting center i in the scheme, and N is the number of supply sorting centers.
  • the objective function of error minimization takes into account the gap between the dismantling enterprise's demand and the actual dispatched sorting center's supply.
  • the absolute value of the expected value of the dismantling enterprise minus the actual quantity of goods transferred in the plan is used as the error.
  • the objective function to minimize the error is:
  • d is the total demand of dismantling enterprises
  • m i is the quality of waste electrical and electronic products supplied by the i-th sorting center.
  • the objective function of satisfaction maximization takes into account the satisfaction index designed in the long-term supply plan, which is a major factor affecting the final transfer plan in the short-term supply.
  • the present invention comprehensively considers the mutual satisfaction between the sorting center and the dismantling enterprise, and designs the objective function of maximizing the satisfaction.
  • the objective function to maximize satisfaction is:
  • a i is the dismantling enterprise's satisfaction with the sorting center
  • b i is the sorting center's satisfaction with the dismantling enterprise.
  • Maximizing the cooperation history takes into account the cooperation history between the sorting center and the dismantling enterprise.
  • the solution enterprise cooperates with the sorting center with the history of cooperation as much as possible, so as to further improve the feasibility of the short-term supply plan, and designs the objective function of maximizing the cooperation history.
  • the objective function for maximizing the cooperation history is:
  • h i is the number of cooperation between sorting center i and the dismantling enterprise.
  • waste electrical and electronic products refer to the short-term supply plan that the goods sent by the sorting center must be the goods designated by the dismantling company.
  • the constraint function of the category constraints of waste electrical and electronic products is:
  • the non-negative constraint of the distribution quantity is to consider that the solution of the intelligent algorithm is uncertain, and it is very likely that the number of shipments in the sorting center will be negative, so constraints are required;
  • the constraint function of the non-negative constraint on the distribution quantity is:
  • the constraint function of the sorting center inventory constraint is to consider that the total delivery of the sorting center cannot exceed the existing inventory of the sorting center; in order to make the solution of the intelligent algorithm have practical application value, it is necessary to limit the total supply to be lower than The total amount of existing inventory, the constraint function of the inventory constraint of the sorting center is:
  • s i is the inventory capacity of dismantling center i.
  • the present invention intelligently analyzes the whole process of electronic solid waste recycling, fully taps the value in historical data, builds a targeted model, uses an intelligent algorithm to solve the model, and improves the information interaction among participating subjects in the whole process of electronic solid waste recycling It reduces the cost of recycling, realizes the informatization and intelligence of electronic solid waste recycling, improves the efficiency and scientificity of electronic solid waste recycling network construction, saves the recycling time of offline recyclers, and reduces the time spent on vehicle transportation. Reduce freight costs, reduce the cost of supply and transfer between the sorting center and the dismantling enterprise, thereby increasing the amount of recycling and satisfaction.
  • Embodiment 2 of the present invention The following is an introduction to the electronic solid waste recycling full-process intelligent analysis system disclosed in Embodiment 2 of the present invention.
  • the electronic solid waste recycling full-process intelligent analysis system described below is the same as the electronic solid waste recycling full-process intelligent Analytical methods can be referred to in correspondence with each other.
  • the present invention also provides an intelligent analysis system for the whole process of electronic solid waste recycling, including:
  • the data acquisition module 10 the data acquisition module 10 is used to acquire the historical data of the electronic solid waste generation amount and the influence solid waste generation amount in the region to be predicted, and perform normalization processing on the historical data, and use the normalization processing Build a multivariate gray model based on historical data, use intelligent learning algorithms to compensate for errors in the predicted values of the multivariate gray model, and obtain the predicted value of electronic solid waste generation in the area to be predicted;
  • the location distribution planning module 20 is used to construct the site selection model of recycling outlets based on the predicted value of electronic solid waste generation, and use the intelligent algorithm to solve the location selection model of recycling outlets to obtain the location distribution of recycling outlets, and according to The location distribution of the recycling outlets constructs a sorting center location model, uses an intelligent algorithm to solve the sorting center location model, and obtains the location distribution of the sorting center;
  • Intelligent dispatching module 30 described intelligent dispatching module 30 is used for constructing the dispatching model of recycling personnel based on the position distribution of described recycling outlets, constructing a dispatching model of recycling vehicles based on the distribution of locations of the recycling outlets and the distribution of sorting centers, based on the distribution of the locations of the recycling outlets
  • the location distribution of the sorting center builds a long-term supply model and a short-term cargo transfer model, and uses intelligent algorithms to solve the recycling personnel scheduling model, recycling vehicle scheduling model, long-term supply model and short-term cargo transfer model to obtain the recycling personnel scheduling plan, recycling Vehicle scheduling plan, long-term supply plan and short-term transfer plan;
  • a visualization module 40 the visualization module 40 is used for visualizing the recycling personnel scheduling plan, recycling vehicle scheduling plan, long-term supply plan and short-term goods transfer plan.
  • the data acquisition module includes an electronic solid waste prediction sub-module, and the electronic solid waste prediction sub-module is used to utilize normalized historical data Construct a multivariate gray model, use the intelligent learning algorithm to compensate the error of the predicted value of the multivariate gray model, and obtain the predicted value of electronic solid waste generation in the area to be predicted, including:
  • represents the delay time
  • r represents the historical sequence number
  • t represents the sequence number
  • a cumulative result of is the cumulative result of Y G (0)
  • Y G (0) is the output prediction sequence of the multivariate gray model
  • both a and u are model parameters
  • a represents the control coefficient
  • u represents the gray action
  • rf represents the time series number to be predicted
  • ⁇ (0) Y (0) - Y G (0)
  • Y G (0) ⁇ Y G (0) ( ⁇ +1),Y G (0) ( ⁇ +2),...,Y G (0) ( ⁇ +r) ⁇ ,
  • ⁇ (0) ⁇ (0) ( ⁇ +1), ⁇ (0) ( ⁇ +2),..., ⁇ (0) ( ⁇ +r) ⁇ ;
  • S15 Use the intelligent learning method to intelligently compensate the prediction error of the multivariate gray model, and obtain the error compensation sequence
  • Y H (0) represents the final intelligent prediction result of electronic solid waste generation
  • the location distribution planning module includes a recycling network construction sub-module, and the recycling network construction sub-module is used to construct a recycling system based on the predicted value of electronic solid waste generation.
  • Site selection model including:
  • a site selection model for recycling sites was established with the goal of minimizing economic costs and maximizing the coverage of solid waste generation centralized points.
  • the location distribution planning module includes a sorting center construction sub-module, and the sorting center construction sub-module is used to construct Sorting center site selection model, including:
  • the electronic solid waste recycling full-process intelligent analysis system of this embodiment is used to implement the aforementioned electronic solid waste recycling full-process intelligent analysis method, so the specific implementation of the system can be seen in the implementation of the electronic solid waste recycling full-process intelligent analysis method above. Therefore, for the specific implementation manners, reference may be made to the descriptions of the corresponding parts of the embodiments, and no further introduction will be made here.
  • the intelligent analysis system for the whole process of electronic solid waste recycling in this embodiment is used to implement the aforementioned intelligent analysis method for the entire process of electronic solid waste recycling, its function corresponds to that of the above method, and will not be repeated here.
  • the embodiments of the present application may be provided as methods, systems, or computer program products. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.
  • computer-usable storage media including but not limited to disk storage, CD-ROM, optical storage, etc.
  • These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing apparatus to operate in a specific manner, such that the instructions stored in the computer-readable memory produce an article of manufacture comprising instruction means, the instructions
  • the device realizes the function specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Human Resources & Organizations (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Operations Research (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • Physics & Mathematics (AREA)
  • Development Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Educational Administration (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Sustainable Development (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

本发明涉及一种电子固废回收全流程智能解析方法,包括获取电子固废产生量历史数据构建多元灰色模型,对多元灰色模型的预测值进行误差补偿,获得电子固废产生量预测值;基于预测值构建回收网点选址模型并对其进行求解,获得回收网点位置分布,根据回收网点位置分布构建分拣中心选址模型并对其进行求解,获得分拣中心位置分布;基于回收网点位置分布和分拣中心位置分布构建回收人员调度模型、回收车辆调度模型、长期供货模型及短期调货模型,分别对其进行求解,获得回收人员调度方案、回收车辆调度方案、长期供货方案及短期调货方案;对方案进行可视化处理。本发明针对电子固废回收全流程进行了智能解析,实现电子固废回收的信息化及智能化。

Description

电子固废回收全流程智能解析方法及系统 技术领域
本发明涉及智能信息交互技术领域,尤其是指一种电子固废回收全流程智能解析方法及系统。
背景技术
联合国大学的报告指出中国是全球第一大电子固废产生国。据统计2019年中国电子固废报废量已达1013万吨,而回收量只有154.6万吨,回收率不足15.3%。由于废旧电子产品中含有大量的重金属及有毒物质,这些没有被回收的废旧电器电子产品不仅会造成环境的污染,同时也导致了资源的浪费。这一现状的原因是电子固废回收流程各环节间的信息交互不畅通甚至没有数字化,而且对已有的数据没有进行深度解析进一步导致回收流程效率低下。因此对废旧电器电子产品的回收全流程进行研究,针对各回收环节解析其数据,将有助于提升废旧电器电子产品的回收效果。
现在的电子固废回收进入互联网+回收模式,流程是消费者线上下单,回收人员上门回收到回收网点,回收网点送到分拣中心进行分拣,最后送到外拆解企业进行破拆。但是现有的电子固废回收流程各环节间的信息交互不畅通甚至没有数字化,而且对已有的数据没有进行深度解析,进一步导致回收流程效率低下,甚至现有的解析基本只针对回收流程中的某环节,忽视各个流程间的数据的相关性,并且也没有充分利用历史数据的价值,导致其解析效果一般,无法提高回收效率。
发明内容
为此,本发明所要解决的技术问题在于克服现有技术存在的缺陷,提供 一种电子固废回收全流程智能解析方法及系统,提高电子固废回收全流程中各参与主体的信息交互程度,降低了回收成本,实现了电子固废回收的信息化及智能化,能够提升电子固废回收网络构建的效率及科学性,节约线下回收人员的回收时间,降低了车辆运输过程中的运费,降低分拣中心与拆解企业间的供货调货成本,从而提高回收量与满意度。
为解决上述技术问题,本发明提供一种电子固废回收全流程智能解析方法,包括:
S1:获取待预测地区的电子固废产生量以及影响固废产生量的历史数据,对所述历史数据进行归一化处理,利用归一化处理后的历史数据构建多元灰色模型,利用智能学习算法对多元灰色模型的预测值进行误差补偿,获得待预测地区的电子固废产生量预测值;
S2:基于电子固废产生量预测值构建回收网点选址模型,利用智能算法对回收网点选址模型进行求解,获得回收网点位置分布,并根据所述回收网点位置分布构建分拣中心选址模型,利用智能算法对分拣中心选址模型进行求解,获得分拣中心位置分布;
S3:基于所述回收网点位置分布构建回收人员调度模型,基于所述回收网点位置分布和所述分拣中心位置分布构建回收车辆调度模型,基于所述分拣中心位置分布构建长期供货模型及短期调货模型,利用智能算法分别对回收人员调度模型、回收车辆调度模型、长期供货模型及短期调货模型进行求解,获得回收人员调度方案、回收车辆调度方案、长期供货方案及短期调货方案;
S4:对所述回收人员调度方案、回收车辆调度方案、长期供货方案及短期调货方案进行可视化处理。
在本发明的一个实施例中,利用归一化处理后的历史数据构建多元灰色 模型,利用智能学习算法对多元灰色模型的预测值进行误差补偿,获得待预测地区的电子固废产生量预测值的方法包括:
S11:利用归一化处理后的历史数据构建多元灰色模型如下:
Figure PCTCN2021131990-appb-000001
其中,τ表示延迟时间,r表示历史时序数,t表示时序数;t=1,2,…r表示时序数,
Figure PCTCN2021131990-appb-000002
Figure PCTCN2021131990-appb-000003
的一次累加结果,
Figure PCTCN2021131990-appb-000004
为Y G (0)的一次累加结果,Y G (0)为多元灰色模型的输出预测序列;a和u均为模型参数,a表示控制系数,u表示灰色作用量;
S12:求解多元灰色模型的响应函数,所述响应函数的计算公式为:
Figure PCTCN2021131990-appb-000005
其中,rt表示待预测的时序数,
Figure PCTCN2021131990-appb-000006
S13:对所述响应函数中的卷积积分进行离散化处理,得到运算结果之后进行一阶累减,得到多元灰色模型的输出预测序列如下:
Y G (0)(τ+t)=Y G (1)(τ+t)-Y G (1)(τ+t-1);
S14:计算多元灰色模型的输出预测序列与实际的电子固废产生量之间的预测误差如下:
σ (0)=Y (0)-Y G (0)
其中,Y G (0)={Y G (0)(τ+1),Y G (0)(τ+2),…,Y G (0)(τ+r)},
σ (0)={σ (0)(τ+1),σ (0)(τ+2),…,σ (0)(τ+r)};
S15:利用智能学习方法对多元灰色模型的预测误差进行智能补偿,得到误差补偿序列;
S16:将误差补偿序列与多元灰色模型的输出预测序列相加,得到最终的电子固废产生量预测结果如下:
Figure PCTCN2021131990-appb-000007
其中,Y H (0)表示最终的电子固废产生量智能预测结果,
Figure PCTCN2021131990-appb-000008
在本发明的一个实施例中,基于电子固废产生量预测值构建回收网点选址模型的方法包括:
对所述电子固废产生量预测值进行归一化处理;
基于归一化处理后的电子固废产生量预测值确定固废产生集中点;
以最小化经济成本和最大化固废产生集中点覆盖率为目标构建回收网点选址模型。
在本发明的一个实施例中,根据所述回收网点位置分布构建分拣中心选址模型的方法包括:
将回收网点和分拣中心作为上层,拆解企业作为下层,基于上下两层的相 互联系和制约构建分拣中心选址模型。
在本发明的一个实施例中,基于所述回收网点位置分布构建回收人员调度模型的方法包括:
将所述回收网点位置分布作为输入,回收人员调度方案作为输出构建回收人员调度模型。
在本发明的一个实施例中,基于所述回收网点位置分布和所述分拣中心位置分布构建回收车辆调度模型包括:
对所述回收网点位置分布和所述分拣中心位置分布进行归一化处理;
将归一化处理后的回收网点位置分布和所述分拣中心位置分布作为输入,回收车辆调度方案作为输出构建回收车辆调度模型。
此外,本发明还提供一种电子固废回收全流程智能解析系统,包括:
数据获取模块,所述数据获取模块用于获取待预测地区的电子固废产生量以及影响固废产生量的历史数据,对所述历史数据进行归一化处理,利用归一化处理后的历史数据构建多元灰色模型,利用智能学习算法对多元灰色模型的预测值进行误差补偿,获得待预测地区的电子固废产生量预测值;
位置分布规划模块,所述位置分布规划模块用于基于电子固废产生量预测值构建回收网点选址模型,利用智能算法对回收网点选址模型进行求解,获得回收网点位置分布,并根据所述回收网点位置分布构建分拣中心选址模型,利用智能算法对分拣中心选址模型进行求解,获得分拣中心位置分布;
智能调度模块,所述智能调度模块用于基于所述回收网点位置分布构建回收人员调度模型,基于所述回收网点位置分布和所述分拣中心位置分布构建回收车辆调度模型,基于所述分拣中心位置分布构建长期供货模型及短期 调货模型,利用智能算法分别对回收人员调度模型、回收车辆调度模型、长期供货模型及短期调货模型进行求解,获得回收人员调度方案、回收车辆调度方案、长期供货方案及短期调货方案;
可视化模块,所述可视化模块用于对所述回收人员调度方案、回收车辆调度方案、长期供货方案及短期调货方案进行可视化处理。
在本发明的一个实施例中,所述数据获取模块包括电子固废预测子模块,所述电子固废预测子模块用于利用归一化处理后的历史数据构建多元灰色模型,利用智能学习算法对多元灰色模型的预测值进行误差补偿,获得待预测地区的电子固废产生量预测值,包括:
S11:利用归一化处理后的历史数据构建多元灰色模型如下:
Figure PCTCN2021131990-appb-000009
其中,τ表示延迟时间,r表示历史时序数,t表示时序数;t=1,2,…r表示时序数,
Figure PCTCN2021131990-appb-000010
Figure PCTCN2021131990-appb-000011
的一次累加结果,
Figure PCTCN2021131990-appb-000012
为Y G (0)的一次累加结果,Y G (0)为多元灰色模型的输出预测序列;a和u均为模型参数,a表示控制系数,u表示灰色作用量;
S12:求解多元灰色模型的响应函数,所述响应函数的计算公式为:
Figure PCTCN2021131990-appb-000013
其中,rt表示待预测的时序数,
Figure PCTCN2021131990-appb-000014
S13:对所述响应函数中的卷积积分进行离散化处理,得到运算结果之后进行一阶累减,得到多元灰色模型的输出预测序列如下:
Y G (0)(τ+t)=Y G (1)(τ+t)-Y G (1)(τ+t-1);
S14:计算多元灰色模型的输出预测序列与实际的电子固废产生量之间的预测误差如下:
σ (0)=Y (0)-Y G (0)
其中,Y G (0)={Y G (0)(τ+1),Y G (0)(τ+2),…,Y G (0)(τ+r)},
σ (0)={σ (0)(τ+1),σ (0)(τ+2),…,σ (0)(τ+r)};
S15:利用智能学习方法对多元灰色模型的预测误差进行智能补偿,得到误差补偿序列;
S16:将误差补偿序列与多元灰色模型的输出预测序列相加,得到最终的电子固废产生量预测结果如下:
Figure PCTCN2021131990-appb-000015
其中,Y H (0)表示最终的电子固废产生量智能预测结果,
Figure PCTCN2021131990-appb-000016
在本发明的一个实施例中,所述位置分布规划模块包括回收网点构建子模块,所述回收网点构建子模块用于基于电子固废产生量预测值构建回收网点选址模型,包括:
对所述电子固废产生量预测值进行归一化处理;
基于归一化处理后的电子固废产生量预测值确定固废产生集中点;
以最小化经济成本和最大化固废产生集中点覆盖率为目标构建回收网点选址模型。
在本发明的一个实施例中,所述位置分布规划模块包括分拣中心构建子模块,所述分拣中心构建子模块用于根据所述回收网点位置分布构建分拣中心选址模型,包括:
将回收网点和分拣中心作为上层,拆解企业作为下层,基于上下两层的相互联系和制约构建分拣中心选址模型。
本发明的上述技术方案相比现有技术具有以下优点:
本发明针对电子固废回收全流程进行了智能解析,充分挖掘历史数据中的价值,并针对性的建立模型,利用智能算法对模型进行求解,提高电子固废回收全流程中各参与主体的信息交互程度,降低了回收成本,实现了电子固废回收的信息化及智能化,能够提升电子固废回收网络构建的效率及科学性,节约线下回收人员的回收时间,降低了车辆运输过程中的运费,降低分拣中心与拆解企业间的供货调货成本,从而提高回收量与满意度。
附图说明
为了使本发明的内容更容易被清楚的理解,下面根据本发明的具体实施例并结合附图,对本发明作进一步详细的说明,其中
图1为本发明实施例一种电子固废回收全流程智能解析方法的流程示意图。
图2为本发明实施例一种电子固废回收全流程智能解析系统的硬件结构示意图。
图3为本发明实施例
附图标记说明如下:10、数据获取模块;20、位置分布规划模块;30、 智能调度模块;40、可视化模块。
具体实施方式
下面结合附图和具体实施例对本发明作进一步说明,以使本领域的技术人员可以更好地理解本发明并能予以实施,但所举实施例不作为对本发明的限定。
实施例一
请参阅图1所示,本实施例提供一种电子固废回收全流程智能解析方法,包括:
S1:获取待预测地区的电子固废产生量以及影响固废产生量的历史数据,对所述历史数据进行归一化处理,利用归一化处理后的历史数据构建多元灰色模型,利用智能学习算法对多元灰色模型的预测值进行误差补偿,获得待预测地区的电子固废产生量预测值;
S2:基于电子固废产生量预测值构建回收网点选址模型,利用智能算法对回收网点选址模型进行求解,获得回收网点位置分布,并根据所述回收网点位置分布构建分拣中心选址模型,利用智能算法对分拣中心选址模型进行求解,获得分拣中心位置分布;
S3:基于所述回收网点位置分布构建回收人员调度模型,基于所述回收网点位置分布和所述分拣中心位置分布构建回收车辆调度模型,基于所述分拣中心位置分布构建长期供货模型及短期调货模型,利用智能算法分别对回收人员调度模型、回收车辆调度模型、长期供货模型及短期调货模型进行求解,获得回收人员调度方案、回收车辆调度方案、长期供货方案及短期调货方案;
S4:对所述回收人员调度方案、回收车辆调度方案、长期供货方案及短 期调货方案进行可视化处理。
在本发明公开的一种电子固废回收全流程智能解析方法中,在步骤S1中,利用归一化处理后的历史数据构建多元灰色模型,利用智能学习算法对多元灰色模型的预测值进行误差补偿,获得待预测地区的电子固废产生量预测值的方法包括:
S11:利用归一化处理后的历史数据构建多元灰色模型如下式(1),具体的,依据电子固废产生量、人口密度以及家庭可支配收入的历史数据构建多元灰色模型,该多元灰色模型的输入为人口密度以及家庭可支配收入序列X i (0)={X i (0)(1),X i (0)(2),…,X i (0)(r)},i=1,2的一次累加结果,表示为X i (1)={X i (1)(1),X i (1)(2),…,X i (1)(r)},i=1,2,其计算公式如下式(2);多元灰色模型的输出预测序列为Y G (0)={Y G (0)(τ+1),Y G (0)(τ+2),…,Y G (0)(τ+r)}的一次累加结果,表示为Y G (1)={Y G (1)(τ+1),Y G (1)(τ+2),…,Y G (1)(τ+r)},其计算公式如下式(2)所示;其中,电子固废的产生量的历史数据为
Y (0)={Y (0)(τ+1),Y (0)(τ+2),…,Y (0)(τ+r)};
Figure PCTCN2021131990-appb-000017
Figure PCTCN2021131990-appb-000018
Figure PCTCN2021131990-appb-000019
其中,τ表示延迟时间,r表示历史时序数,t表示时序数;t=1,2,…r表示 时序数,
Figure PCTCN2021131990-appb-000020
Figure PCTCN2021131990-appb-000021
的一次累加结果,
Figure PCTCN2021131990-appb-000022
为Y G (0)的一次累加结果,Y G (0)为多元灰色模型的输出预测序列;a和u均为模型参数,a表示控制系数,u表示灰色作用量;
S12:求解多元灰色模型的响应函数,所述响应函数的计算公式为式(3):
Figure PCTCN2021131990-appb-000023
其中,rf表示待预测的时序数,
Figure PCTCN2021131990-appb-000024
S13:对所述响应函数中的卷积积分进行离散化处理,得到运算结果之后进行一阶累减,得到多元灰色模型的输出预测序列如下式(4):
Y G (0)(τ+t)=Y G (1)(τ+t)-    (4)
S14:计算多元灰色模型的输出预测序列与实际的电子固废产生量之间的预测误差σ (0)={σ (0)(τ+1),σ (0)(τ+2),…,σ (0)(τ+r)}如下式(5):
σ (0)=Y (0)-Y G (0)     (5)
其中,Y G (0)={Y G (0)(τ+1),Y G (0)(τ+2),…,Y G (0)(τ+r)},
σ (0)={σ (0)(τ+1),σ (0)(τ+2),…,σ (0)(τ+r)};
S15:利用智能学习方法对多元灰色模型的预测误差进行智能补偿,得到误差补偿序列;
S16:将误差补偿序列与多元灰色模型的输出预测序列相加,得到最终的 电子固废产生量预测结果Y H (0)如下式(6):
Figure PCTCN2021131990-appb-000025
其中,Y H (0)表示最终的电子固废产生量智能预测结果,
Figure PCTCN2021131990-appb-000026
在上述步骤S15中,所述智能学习方法可以为:神经网络、支持向量机、专家系统、模糊系统、深度学习、强化学习。
在本发明公开的一种电子固废回收全流程智能解析方法中,在步骤S2中,基于电子固废产生量预测值构建回收网点选址模型的方法包括:
获取待构建城市各区的电子固废产生量预测值以及经济指标数据,并对其进行归一化处理;
基于归一化处理后的电子固废产生量预测值确定固废产生集中点,具体的,可以取固废产生集中地区的几何中心作为固废产生集中点;
以最小化经济成本和最大化固废产生集中点覆盖率为目标构建回收网点选址模型。
在上述步骤S2中,在构建回收网点选址模型后,利用智能算法对回收网点选址模型进行求解,得到回收网点位置分布,并根据各回收网点所覆盖的固废产生量预测值,按照一定的系数求得回收网点所配置的人员数量及车辆等信息。
在上述步骤S2中,电子固废产生量预测值由步骤S1预测所得,经济指标数据包括:建筑成本、运营成本、运输费用及货损费用。
在本发明公开的一种电子固废回收全流程智能解析方法中,在步骤S2中,根据所述回收网点位置分布构建分拣中心选址模型的方法包括:
将回收网点和分拣中心作为上层,拆解企业作为下层,基于上下两层的相互联系和制约构建分拣中心选址模型。
在上述步骤S2中,在构建分拣中心选址模型后,利用智能算法对分拣中心选址模型进行求解,得到分拣中心位置分布,并根据分拣中心所覆盖回收网点回收量,按照一定的系数求得回收网点所配置的人员数量及车辆等信息。
在上述步骤S2中,基于上下两层的相互联系和制约是上层主要考虑:经济指标及固废产生量预测结果等,下层主要考虑拆解企业的满意度与经济指标等。其中上层经济指标包括建筑成本、运营成本、运输费用及货损费用,下层经济指标包括交易价格、运输费用及货损费用。
在上述步骤S2中,上层主要考虑了决策层的成本控制问题,即从居民回收到分拣中心的上层,主要考虑的成本包括建筑成本、运营成本以及可变投资成本,其中可变投资成本主要包含运输费用,货物损害费用等。因此可得到上层模型的表达式如下:
Figure PCTCN2021131990-appb-000027
Figure PCTCN2021131990-appb-000028
Figure PCTCN2021131990-appb-000029
其约束条件为:
Figure PCTCN2021131990-appb-000030
Figure PCTCN2021131990-appb-000031
V i≤MaxV i      (12)
y i={0,1,}              (13)
其中,约束式(10)表示所有分拣中心的投资成本之和有上限,约束式(11)表示至少有一个备选分拣中心被选中,约束式(12)表示分拣中心的库存必须小于最大库存;约束式(13)保证该上层模型的解为0-1型。
在上述步骤S2中,下层主要考虑本发明的后端即拆解企业是第三方企业,无法对其进行控制。因此我们主要考虑对拆解企业做到在保证自身成本最低的目标函数下,对供货的拆解企业服务满意度最高。因此需要考虑的主要有价格、距离以及货物损害情况等。因此下层模型的表达式如下:
Figure PCTCN2021131990-appb-000032
Figure PCTCN2021131990-appb-000033
其约束条件如下:
Figure PCTCN2021131990-appb-000034
x ij≤y j         (17)
x ij={0,1}            (18)
d ij≤D         (19)
Figure PCTCN2021131990-appb-000035
j∈1,2,...,m          (21)
其中,约束式(16)保证下层模型得到的解为0-1型数值,约束式(17)表示分拣中心与拆解企业合作建立在该地区建有分拣中心,约束式(18)表示分拣中心与拆解企业合作的模式,1代表合作,否则为0,约束式(19)表示分拣中心到拆解企业的距离不能超过最大距离。
在上述步骤S2中,对于模型的求解,尤其是针对于小规模的模型而言,变量个数较少,上层模型可以通过枚举法如分支定界法列出所有决策方案,并作为下层决策变量的解输入到下层模型,然后运用MATLAB函数工具箱求解,再将计算出的(Q,B)返回到上层目标函数,计算出对应决策方案的成本,筛选出成本最低的选址方案。
在本发明公开的一种电子固废回收全流程智能解析方法中,在步骤S3中,基于所述回收网点位置分布构建回收人员调度模型的方法包括:
将所述回收网点位置分布作为输入,回收人员调度方案作为输出构建回收人员调度模型。
在上述步骤S3中,回收人员调度模型的构建还与订单信息相关,例如具体如下:
获取线上订单信息及线下回收人员状态信息;在新增线上订单时,利用线上订单信息计算运输时间、拆卸时间及搬运时间等,合计为该订单预估所需回收时间,同时利用线下回收人员状态信息筛选出能够服务的回收人员;在筛选后的回收人员中,按照当前单数少、距离客户近及熟练程度高等指标综合排序;利用排序信息为排序第一的回收人员增加回收任务,更新回收人员服务时间列表;在用户删除订单或回收人员上门取件失败时,将相关回收人员该时间段时间列表置零,并更新回收人员服务时间列表;构建以回收人员服务时间列表为输入,以回收人员调度方案为输出的回收人员调度模型。
在上述步骤S3中,所述线上订单信息包括货品种类、有无电梯、货物体积、货品重量、回收地点、是否分批、特殊要求等;线下回收人员状态信息包括回收人员服务时间列表、回收人员当前位置、天气变化、交通状况、路况优劣、车辆性能等。
在上述步骤S3中,在一个优选的方案中,假设回收网点有K辆运输车辆,K={1,2,...,m}为车辆集,假设所有车辆均为一种类型,最大载重量为q;该回收网点向n个客户回收废弃电器电子产品,V={0,1,2,...n}表示客户集,A={d ij|i,j∈V,i≠j}表示两客户间的直线距离;第i个客户货品存量为g i,且g i≤q,从而保证一个客户仅由一辆车配送。车辆路径优化问题的目标是合理安排运输车辆并确定每辆车的运输路线,从而保证运输路程最短。另外,给出以下约束条件:
每个客户仅由一辆车进行配送:
Figure PCTCN2021131990-appb-000036
其中,y ik表示第k辆车负责第i个客户的运输,i表示第i个客户,i∈V;k表示第k辆车,k∈K。
车辆k从客户i行驶至客户j:
Figure PCTCN2021131990-appb-000037
其中,x ijk表示车辆k从客户i行驶至客户j。
每条线路总运输货物重量不超过车辆的最大载重量:
Figure PCTCN2021131990-appb-000038
其中,g i表示第i个客户货品存量,
Figure PCTCN2021131990-appb-000039
表示每条线路的总配送量。
上述车辆路径优化问题的目标函数为:
Figure PCTCN2021131990-appb-000040
其中,d ij表示客户i到客户j的距离。
基于上述内容的描述,本实施例将算法流程描述如下:
步骤1:在各城市地理位置已知的前提下,计算回收网点到客户以及不同客户之间的最短距离。车辆路径优化问题的解为一组遍历所有客户的路径。设两客户的坐标为(x i,y i)、(x j,y j),其中,x i表示客户i位置的横坐标,y i表示纵坐标,计算其距离c ij
Figure PCTCN2021131990-appb-000041
步骤2:计算节约里程,具体包括以下步骤:
步骤2.1:计算节约里程数P ij:
p ij=c i,0+c 0,j-c ij          (27)
其中,P ij表示客户i与客户j的节约里程数,c ij表示客户i到客户j的距离,c i,0表示客户i到回收网点之间的直线距离,c 0,j表示客户j到回收网点之间的直线距离。
步骤2.2:对节约里程数降序排序并存储在集合S中,S={P ij|P ij>0},若S为空集则计算结束,否则,对S中的P ij所涉及的客户点s i、s j进行判断,判断其是否满足以下条件:
1)s i、s j均不在已构成的路径上;
2)s i和s j一个为路径起点,一个为路径终点;
3)s i不在路径上,s j为路径起点;
4)s i不在路径上,s j为路径终点;
若满足上述条件之一,则进行步骤3,否则,从S中删掉当前P ij
步骤3:连接s i和s j为一条路径,判断包含s i和s j的路径上所有客户货物总量是否满足车辆最大载重量的要求,若满足,则将其记作v l(l=1,2,…,n),并加入路径集合中,判断下一组P ij中涉及到的客户点s i、s j是否能进行连接为一条路径。
步骤4:重复步骤2、3中的内容,直至S为空集结束。
在本发明公开的一种电子固废回收全流程智能解析方法中,在步骤S3中,基于所述回收网点位置分布和所述分拣中心位置分布构建回收车辆调度模型 包括:
获取回收网点位置分布、分拣中心位置分布、当前各回收网点库存容量及分拣中心车辆信息,并进行归一化处理;
将归一化处理后的回收网点位置分布和所述分拣中心位置分布作为输入,回收车辆调度方案作为输出构建回收车辆调度模型。
在上述步骤S3中,所述分拣中心车辆信息包括:耗油量、承载能力、最高时速、车辆类型及最大行驶距离等。
在上述步骤S3中,在一个优选的方案中,假设有n个回收网点,m辆配送车辆,每个回收网点需求量为g i(i=1,2,…,n),配送车辆类型完全一致,每辆车的最大载重量为q,最大装载量为V,网点i到网点j的距离为d ij,0表示配送中心,配送中心到回收网点的距离为d 0j(j=1,2,…,n),
Figure PCTCN2021131990-appb-000042
表示配送线路k所包含的回收网点数,
Figure PCTCN2021131990-appb-000043
表示配送线路k的行驶里程。因为每辆车负责一条线路上所有回收网点的配送任务,所以要求每条线路回收网点的需求总量不超过每辆车的最大装载量。因此,模型建立如下:
模型需满足的约束条件包括:
条件1:每条线路的货物量(重量、体积)不能超过每辆车的最大载重,即:
Figure PCTCN2021131990-appb-000044
其中,g i表示第i个客户货品存量,vi表示第i个客户货品体积,
Figure PCTCN2021131990-appb-000045
表示每条线路配送货物总重量,
Figure PCTCN2021131990-appb-000046
表示每条线路配送货物总体积。
条件2:如果车辆k从网点i行驶至网点j则为1,否则为0:
Figure PCTCN2021131990-appb-000047
其中,x ijk表示车辆k从客户i行驶至客户j。
条件3:如果网点i的配送任务由车辆k完成则为1,否则为0:
Figure PCTCN2021131990-appb-000048
其中,y ik表示第k辆车负责第i个客户的运输,i表示第i个客户,i∈V;k表示第k辆车,k∈K。
目标函数为:
目标1:所有配送线路的总送货里程数最小,即:
Figure PCTCN2021131990-appb-000049
目标2:一次派出的配送车辆数最少:
Figure PCTCN2021131990-appb-000050
因此,废旧电器电子产品车辆路径优化问题可描述为如下形式:
Figure PCTCN2021131990-appb-000051
基于上述阐述的内容,使用GA-MOEAD算法求解式(33)的流程描述如下:
步骤1:随机产生种群(x 1,x 2,...,x n),并判断当前种群中的个体是否满足最大载重量和最少车辆的要求;
步骤2:计算目标函数值f 1(x i)和f 2(x i)并产生权值向量;
步骤3:计算任意两个权值向量之间的欧式距离,找出T个最近的权值向量。令B(i)={i 1,i 2,...,i T},其中i=1,2,…,n,
Figure PCTCN2021131990-appb-000052
是λ i最近的T个权值向量;
步骤4:从B(i)中随机选择两个解,执行交叉操作产生子代y,再对y执行变异操作产生y';
步骤5:判断y'是否满足限重要求,如果不满足,删除y'并返回步骤4;
步骤6:储存子代并更新外部种群(简称EP);
步骤7:判断是否达到终止条件,未达到则返回步骤5。
在本发明公开的一种电子固废回收全流程智能解析方法中,在步骤S3中,基于所述分拣中心位置分布构建长期供货模型的方法包括:
获取拆解企业及分拣中心的评价数据以及所有拆解企业和分拣中心的地理信息及体量数据,并进行归一化处理;利用模糊方法处理上述评价数据,获得双方的满意度矩阵;利用双方的满意度矩阵,构建以双方满意度最大为目标,结合适当约束的长期供货模型。
在上述步骤S3中,所述评价数据为货物种类、拆解范围、收货价格、分拣中心交易信誉、品牌范围、是否自备车辆、供货价格、发货速度及拆解企业交易信誉等。
在上述步骤S3中,在废旧电子产品回收中,分拣中心只能与一个拆解中心进行匹配,而拆解中心理论上没有分拣中心的数量限制,即为一对多的双边匹配。设所有的分拣中心集合为U={U 1,U 2,…U m},拆解中心的集合为V={V 1,V 2,…V n},其中m和n分别表示分拣中心U和拆解中心V的个数,并且m≥2,n≥2。U i表示分拣中心集合U的第i个个体(i=1,2,…,m),V j表示拆解中心集合V的第j个个体(j=1,2,…,n)。同时双方的评价体系包括:分拣中心对拆解中心的评价体系
Figure PCTCN2021131990-appb-000053
其中E b表示第b个评价指标(b=0,1,…,g),包括语言评价指标
Figure PCTCN2021131990-appb-000054
0-1评价指标
Figure PCTCN2021131990-appb-000055
以及区间型评价指标等;同样拆解中心对分拣中心的指标体系为:
Figure PCTCN2021131990-appb-000056
其中A p表示第p个评价指标 (p=0,1,…,h),同样包括语言评价指标
Figure PCTCN2021131990-appb-000057
0-1评价指标
Figure PCTCN2021131990-appb-000058
以及区间型评价指标等。为方便表达,记M={1,2,…,m},N={1,2,…,n},G={1,2,…,g),H={1,2,…,h}。
下面给出长期供货的定义:
定义:长期供货中有映射ξ:U∪V→U∪V,且
Figure PCTCN2021131990-appb-000059
都满足以下条件:
1)ξ(U i)∈V;
2)ξ(V j)∈U∪V j
3)ξ(U i)=V j当且仅当ξ(V j)=U i
4)若存在ξ(V j)=V i,则V j未完成匹配。
在长期供货中,ξ(U i)≠V j表示U i和V j未达成匹配,(U i,V j)表示U i和V j达成匹配,也称作ξ匹配对;若(U i,V j)为ξ匹配对,则(V j,U i)也为ξ匹配对。ξ(V j)=V i表示V j未达成匹配,记ξ匹配。因此,ξ匹配可表示为ξ=ξ c∪ξ u,其中ξ c={(U i,U C(i))|i=1,2,..,m},ξ U={(V j,V j)|{j=1,2,...,n}\{C(1),C(2),...,C(m)}},C(i)∈N,V C(i)=ξ(U i)并 且
Figure PCTCN2021131990-appb-000060
都有C(i)≠C(k)。
根据上述分析,长期供货决策过程中,主要是考虑分拣中心和拆解中心的评价指标体系给出相对应的综合满意度值,最终根据满意度值进行匹配。
由以上分析,已知分拣中心U对拆解中心V需要考虑的评价体系为
Figure PCTCN2021131990-appb-000061
同样地,拆解中心V对分拣中心U的指标体系为
Figure PCTCN2021131990-appb-000062
根据实际情况,评价体系主要涉及四种类型指标,其中
Figure PCTCN2021131990-appb-000063
为S 1类0-1指标,
Figure PCTCN2021131990-appb-000064
为S 2类语言型指标,同理有S 3区间型指标和S 4一般数值型指标。
针对S 1类0-1型指标如手机能否开机、有无维修记录等,需要拆解中心V j对分拣中心的第p个指标设置期望值
Figure PCTCN2021131990-appb-000065
为分拣中心产品的实际值。若
Figure PCTCN2021131990-appb-000066
则该评价指标值
Figure PCTCN2021131990-appb-000067
为1,反之则为0。拆解中心根据实际情况得到分拣中心对拆解中心的0-1评价矩阵
Figure PCTCN2021131990-appb-000068
以及拆解中心对分拣中心的评价矩阵
Figure PCTCN2021131990-appb-000069
关于S 1类满意度计算公式:
Figure PCTCN2021131990-appb-000070
由式(34)可得到分拣中心对拆解中心的满意度矩阵
Figure PCTCN2021131990-appb-000071
和拆解中心对分 拣中心的满意度矩阵
Figure PCTCN2021131990-appb-000072
S 2类语言评价指标如分拣中心信誉度、产品描述符合度等,S 2语言评价信息一般由奇数个元素构成的有序集s={s 1,s 2,…,s t},分拣中心根据实际情况对产品进行描述得到评价指标s q,将其转化为三角模糊函数f=(f 1,f 2,f 3),可得到S 2类语言评价矩阵
Figure PCTCN2021131990-appb-000073
Figure PCTCN2021131990-appb-000074
相应的转化公式:
Figure PCTCN2021131990-appb-000075
最后根据分拣中心和拆解中心的偏好设定
Figure PCTCN2021131990-appb-000076
Figure PCTCN2021131990-appb-000077
得到相对应的三角模糊值
Figure PCTCN2021131990-appb-000078
得到对应的满意度矩阵
Figure PCTCN2021131990-appb-000079
Figure PCTCN2021131990-appb-000080
则关于S 2类语言评价满意度值计算公式:
Figure PCTCN2021131990-appb-000081
同样,S 3类区间型评价指标如产品价格,首先基于分拣中心对于产品的属性参数根据定价系统获得一定的参考定价
Figure PCTCN2021131990-appb-000082
同时分拣中心基于给出自己的期望价格
Figure PCTCN2021131990-appb-000083
x 1≤x 2且x 1,x 2∈R,得到分拣中心对拆解中心的区间满意度值矩阵
Figure PCTCN2021131990-appb-000084
和拆解中心对分拣中心的区间满意度矩阵
Figure PCTCN2021131990-appb-000085
则关于S 3类区间型评价满意度值计算公式:
Figure PCTCN2021131990-appb-000086
根据公式(35)-(37)可以分别得出分拣中心对拆解中心以及拆解中心对分拣中心的满意度值矩阵
Figure PCTCN2021131990-appb-000087
Figure PCTCN2021131990-appb-000088
当匹配双方主体有硬性考量指标时,已有匹配算法并未实现对其精确筛选;同样,匹配双方对于期望指标更满意的主体未实现优先匹配。因此,本发明加入硬约束和软约束。硬约束是指一方指标属性值必须达到另一方的要求,即当主体U i对于主体V中的指标A p有硬性要求,若主体V j未达到U i的硬约束,则在满意度计算时赋极大负值-K i,使其与无法与主体U i实现匹配;软约束是指一方所提供指标属性值尽量能达到另一方的期望,即主体U i对主体V中的指标A p满意度值期望越高越好。计算公式如下:
Figure PCTCN2021131990-appb-000089
Figure PCTCN2021131990-appb-000090
其中式(38)中,
Figure PCTCN2021131990-appb-000091
表示主体U i对主体V中的指标A p有硬约束,并赋负值K i,且K i足够大。式(39)中,
Figure PCTCN2021131990-appb-000092
表示主体U i对主体V中的指标A p有软约束,K i表示A p指标下的赋值,
Figure PCTCN2021131990-appb-000093
表示该指标根据式35-37算得的值。
由于每一个分拣中心对产品回收的需求大相径庭,给予统一的权重的分配显然不合理,因此本文针对废旧电子回收过程设计随机偏好序权重分配。而拆解中心对分拣中心的指标偏好相对固定,一般按照主观经验设定给出
Figure PCTCN2021131990-appb-000094
具体如图3所示,其中指标序依次从左至右,重要程度依次降低,E a,E b,E c为指标体系E中的指标,给定指标序列1~g分别给定权重分配
Figure PCTCN2021131990-appb-000095
并使得
Figure PCTCN2021131990-appb-000096
指标序值表根据分拣中心偏好设定,并自动更新相对应的满意度矩阵
Figure PCTCN2021131990-appb-000097
Figure PCTCN2021131990-appb-000098
根据式(35)-(37)对指标的处理方法,公式(34)-(39)计算得到的值为评价指标与期望之间的差距。根据图3权重的分配可得到差异矩阵
Figure PCTCN2021131990-appb-000099
以及
Figure PCTCN2021131990-appb-000100
其计算公式为:
Figure PCTCN2021131990-appb-000101
Figure PCTCN2021131990-appb-000102
其中,
Figure PCTCN2021131990-appb-000103
来自图3的指标序值表,
Figure PCTCN2021131990-appb-000104
由决策者通过经验主观给定,最后得到的差异矩阵表示分拣中心或网店的期望与实际值之间的差距,因此可以利用转化公式得到分拣中心对拆解中心的满意度值矩阵
Figure PCTCN2021131990-appb-000105
以及拆解中心对分拣中心的满意度值矩阵
Figure PCTCN2021131990-appb-000106
Figure PCTCN2021131990-appb-000107
Figure PCTCN2021131990-appb-000108
在废旧电子产品回收流程中,分拣中心和拆解中心主要关注的指标信息和属性按照其特定的序值偏好进行计算,并按照本发明提出的满意度值计算方式进行改进,最终得到关于分拣中心和拆解中心的满意度值矩阵。因此为确定分拣中心和拆解中心之间的多目标优化模型,可根据式(42)和(43) 得到的满意度值矩阵可分别建立基于满意度值的多目标优化模型:
Figure PCTCN2021131990-appb-000109
Figure PCTCN2021131990-appb-000110
Figure PCTCN2021131990-appb-000111
x ij=0或1,j=1,2,…,m;j=1,2,…,n    (47)
其中,式(44)和(45)分别表示使分拣中心和拆解中心达到最大满意度值的目标函数;同时x ij的值即表示了分拣中心和拆解中心的匹配结果。约束条件46表示分拣中心最多仅能匹配一个拆解中心,式(47)限定x ij即匹配结果只能为0或1,当x ij=1时表示分拣中心i与拆解中心j达成匹配,反之即x ij=0表示未达成匹配。
通过建立式(44)-(47)的多目标优化模型实现了对分拣中心和拆解中心满意度最大化的目标,求解得到的结果即为分拣中心和拆解中心整体满意度值最大。而针对此多目标模型,本文使用较为广泛的线性加权的求解方式。首先通过主观经验给予不同的加权和的权重参数ω 1,ω 2,进而可实现总满意度值Z=ω 1Z 12Z 2最大:
Figure PCTCN2021131990-appb-000112
Figure PCTCN2021131990-appb-000113
x ij=0或1,j=1,2,…,m;j=1,2,…,n    (50)
其中,
Figure PCTCN2021131990-appb-000114
ω 1,ω 2的值也显示了本次匹配中对于分拣中心和拆解中心和重视程度,理论情况下通常取ω 1=ω 2=0.5。但具体取值以实际情况而论。
在本发明公开的一种电子固废回收全流程智能解析方法中,在步骤S3中,基于所述分拣中心位置分布构建短期调货模型的方法包括:
获取当前各分拣中心库存容量,拆解企业所需货物种类及数量及历史交易数据、长期供货子模块中的满意度矩阵、所有拆解企业和分拣中心的地理信息及体量数据,并进行归一化处理;利用拆解企业及分拣中心历史交易数据,获取各分拣中心与拆解企业的合作次数;利用拆解企业和分拣中心的地理信息及货物数量计算出运费;构建以运费最小化、调货误差最小化、满意度最大化为目标,结合相应约束的短期调货模型。
在上述步骤S3中,首先需要分析供货方案的期望,在拆解企业的短期供货问题中,供需双方的主要成本来自运输废旧电器电子产品产生的运费,同时短期供货方案应尽可能的接近拆解企业的需求值,此外如长期供货中所述拆解企业和分拣中心间的满意度也会影响短期供货方案的可行性,在实际执行情况中拆解企业也倾向于同已有合作历史的分拣中心再次合作。综上所述,本发明设置四个目标,分别为:运费最小化,误差最小化,双方满意度最大化及合作历史最大化。
为了使生成的供货方案具有可行性,还需要考虑实际供货时的约束条件,在确定了短期供货模型的目标与约束后,本发明对其进行了数学化表达,具体的表达式如下所示:
运费最小化目标函数综合考虑了在分拣中心与拆解企业进行交易的过程中的运输距离和运输货物重量,该目标函数的目的是:在总运输货物重量一 定的前提下,为靠近的拆解企业分配较多的货物,为较远的拆解企业分配较少的货物以降低运费。该目标值越大说明运费得到了更好的节约,反之则运费较为高昂。其中运费最小化的目标函数为:
Figure PCTCN2021131990-appb-000115
其中,caf i为方案中分拣中心i计算得到的运费,N为供货分拣中心的数量。
误差最小化的目标函数考虑了拆解企业需求与实际调集到的分拣中心供货总和之间的差距。为了使最后的短期方案所调配的废旧电器电子产品尽可能的接近拆解企业的需求值,这里用拆解企业期望值减去方案实际的调货数量的绝对值作为误差。该目标值越小,说明分拣中心供货量与拆解企业需求量越接近。其中误差最小化的目标函数为:
Figure PCTCN2021131990-appb-000116
其中,d为拆解企业的需求总量,m i为第i个分拣中心供出的废旧电器电子产品质量。
满意度最大化的目标函数考虑了长期供货方案中设计的满意度指标,这一指标在短期供货中是影响最终调货方案的一大因素,为了使得最终的短期供货方案能够更具备可行性,本发明综合考虑分拣中心与拆解企业互相的满意度,设计了满意度最大化目标函数。其中满意度最大化的目标函数为:
Figure PCTCN2021131990-appb-000117
其中,a i为拆解企业对分拣中心的满意度;b i为分拣中心对拆解企业的满意度。
最大化合作历史考虑了分拣中心与拆解企业之间的合作历史,通过爱博绿分拣仓储云系统中的订单模块,可以统计出各分拣中心与拆解企业的合作历史,为了拆解企业尽可能与有过合作历史的分拣中心再次合作,从而进一步提高短期供货方案的可行性,设计了合作历史最大化目标函数。其中合作历史最大化的目标函数为:
Figure PCTCN2021131990-appb-000118
其中,h i为分拣中心i与该拆解企业的合作次数。
废旧电器电子产品类约束指的是短期供货方案中分拣中心发出的货物需为拆解企业指定的货物。其中废旧电器电子产品品类约束的约束函数为:
Figure PCTCN2021131990-appb-000119
配货数量非负约束是考虑到智能算法的解具有不确定性,很有可能会出现分拣中心发货数为负数的情况,因此需要约束;
所述配货数量非负约束的约束函数为:
g 2=m i≥0        (56)。
分拣中心库存约束的约束函数是考虑到分拣中心发货总量不能超过分拣中心的现有库存;为了使智能算法的解具有实际应用价值,需限制总供货量低于分拣中心现有库存总量,所述分拣中心库存约束的约束函数为:
g 3=m i≤s i      (57)
其中,s i为拆解中心i的库存容量。
本发明针对电子固废回收全流程进行智能解析,充分挖掘历史数据中的 价值,并针对性的建立模型,利用智能算法对模型进行求解,提高电子固废回收全流程中各参与主体的信息交互程度,降低了回收成本,实现了电子固废回收的信息化及智能化,能够提升电子固废回收网络构建的效率及科学性,节约线下回收人员的回收时间,降低了车辆运输过程中的运费,降低分拣中心与拆解企业间的供货调货成本,从而提高回收量与满意度。
实施例二
下面对本发明实施例二公开的一种电子固废回收全流程智能解析系统进行介绍,下文描述的一种电子固废回收全流程智能解析系统与上文描述的一种电子固废回收全流程智能解析方法可相互对应参照。
请参阅图2所示,本发明还提供一种电子固废回收全流程智能解析系统,包括:
数据获取模块10,所述数据获取模块10用于获取待预测地区的电子固废产生量以及影响固废产生量的历史数据,对所述历史数据进行归一化处理,利用归一化处理后的历史数据构建多元灰色模型,利用智能学习算法对多元灰色模型的预测值进行误差补偿,获得待预测地区的电子固废产生量预测值;
位置分布规划模块20,所述位置分布规划模块20用于基于电子固废产生量预测值构建回收网点选址模型,利用智能算法对回收网点选址模型进行求解,获得回收网点位置分布,并根据所述回收网点位置分布构建分拣中心选址模型,利用智能算法对分拣中心选址模型进行求解,获得分拣中心位置分布;
智能调度模块30,所述智能调度模块30用于基于所述回收网点位置分布构建回收人员调度模型,基于所述回收网点位置分布和所述分拣中心位置分布构建回收车辆调度模型,基于所述分拣中心位置分布构建长期供货模型及短期调货模型,利用智能算法分别对回收人员调度模型、回收车辆调度模型、 长期供货模型及短期调货模型进行求解,获得回收人员调度方案、回收车辆调度方案、长期供货方案及短期调货方案;
可视化模块40,所述可视化模块40用于对所述回收人员调度方案、回收车辆调度方案、长期供货方案及短期调货方案进行可视化处理。
在本发明公开的一种电子固废回收全流程智能解析系统中,所述数据获取模块包括电子固废预测子模块,所述电子固废预测子模块用于利用归一化处理后的历史数据构建多元灰色模型,利用智能学习算法对多元灰色模型的预测值进行误差补偿,获得待预测地区的电子固废产生量预测值,包括:
S11:利用归一化处理后的历史数据构建多元灰色模型如下:
Figure PCTCN2021131990-appb-000120
其中,τ表示延迟时间,r表示历史时序数,t表示时序数;t=1,2,…r表示时序数,
Figure PCTCN2021131990-appb-000121
Figure PCTCN2021131990-appb-000122
的一次累加结果,
Figure PCTCN2021131990-appb-000123
为Y G (0)的一次累加结果,Y G (0)为多元灰色模型的输出预测序列;a和u均为模型参数,a表示控制系数,u表示灰色作用量;
S12:求解多元灰色模型的响应函数,所述响应函数的计算公式为:
Figure PCTCN2021131990-appb-000124
其中,rf表示待预测的时序数,
Figure PCTCN2021131990-appb-000125
S13:对所述响应函数中的卷积积分进行离散化处理,得到运算结果之后 进行一阶累减,得到多元灰色模型的输出预测序列如下:
Y G (0)(τ+t)=Y G (1)(τ+t)-Y G (1)(τ+t-1);
S14:计算多元灰色模型的输出预测序列与实际的电子固废产生量之间的预测误差如下:
σ (0)=Y (0)-Y G (0)
其中,Y G (0)={Y G (0)(τ+1),Y G (0)(τ+2),…,Y G (0)(τ+r)},
σ (0)={σ (0)(τ+1),σ (0)(τ+2),…,σ (0)(τ+r)};
S15:利用智能学习方法对多元灰色模型的预测误差进行智能补偿,得到误差补偿序列;
S16:将误差补偿序列与多元灰色模型的输出预测序列相加,得到最终的电子固废产生量预测结果如下:
Figure PCTCN2021131990-appb-000126
其中,Y H (0)表示最终的电子固废产生量智能预测结果,
Figure PCTCN2021131990-appb-000127
在本发明公开的一种电子固废回收全流程智能解析系统中,所述位置分布规划模块包括回收网点构建子模块,所述回收网点构建子模块用于基于电子固废产生量预测值构建回收网点选址模型,包括:
对所述电子固废产生量预测值进行归一化处理;
基于归一化处理后的电子固废产生量预测值确定固废产生集中点;
以最小化经济成本和最大化固废产生集中点覆盖率为目标构建回收网点选址模型。
在本发明公开的一种电子固废回收全流程智能解析系统中,所述位置分布规划模块包括分拣中心构建子模块,所述分拣中心构建子模块用于根据所述回收网点位置分布构建分拣中心选址模型,包括:
将回收网点和分拣中心作为上层,拆解企业作为下层,基于上下两层的相互联系和制约构建分拣中心选址模型。
本实施例的电子固废回收全流程智能解析系统用于实现前述的电子固废回收全流程智能解析方法,因此该系统的具体实施方式可见前文中的电子固废回收全流程智能解析方法的实施例部分,所以,其具体实施方式可以参照相应的各个部分实施例的描述,在此不再展开介绍。
另外,由于本实施例的电子固废回收全流程智能解析系统用于实现前述的电子固废回收全流程智能解析方法,因此其作用与上述方法的作用相对应,这里不再赘述。
本领域内的技术人员应明白,本申请的实施例可提供为方法、系统、或计算机程序产品。因此,本申请可采用完全硬件实施例、完全软件实施例、或结合软件和硬件方面的实施例的形式。而且,本申请可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
本申请是参照根据本申请实施例的方法、设备(系统)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器,使 得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中,使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品,该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上,使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理,从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。
显然,上述实施例仅仅是为清楚地说明所作的举例,并非对实施方式的限定。对于所属领域的普通技术人员来说,在上述说明的基础上还可以做出其它不同形式变化或变动。这里无需也无法对所有的实施方式予以穷举。而由此所引申出的显而易见的变化或变动仍处于本发明创造的保护范围之中。

Claims (10)

  1. 一种电子固废回收全流程智能解析方法,其特征在于,包括:
    S1:获取待预测地区的电子固废产生量以及影响固废产生量的历史数据,对所述历史数据进行归一化处理,利用归一化处理后的历史数据构建多元灰色模型,利用智能学习算法对多元灰色模型的预测值进行误差补偿,获得待预测地区的电子固废产生量预测值;
    S2:基于电子固废产生量预测值构建回收网点选址模型,利用智能算法对回收网点选址模型进行求解,获得回收网点位置分布,并根据所述回收网点位置分布构建分拣中心选址模型,利用智能算法对分拣中心选址模型进行求解,获得分拣中心位置分布;
    S3:基于所述回收网点位置分布构建回收人员调度模型,基于所述回收网点位置分布和所述分拣中心位置分布构建回收车辆调度模型,基于所述分拣中心位置分布构建长期供货模型及短期调货模型,利用智能算法分别对回收人员调度模型、回收车辆调度模型、长期供货模型及短期调货模型进行求解,获得回收人员调度方案、回收车辆调度方案、长期供货方案及短期调货方案;
    S4:对所述回收人员调度方案、回收车辆调度方案、长期供货方案及短期调货方案进行可视化处理。
  2. 根据权利要求1所述的电子固废回收全流程智能解析方法,其特征在于:利用归一化处理后的历史数据构建多元灰色模型,利用智能学习算法对多元灰色模型的预测值进行误差补偿,获得待预测地区的电子固废产生量预测值的方法包括:
    S11:利用归一化处理后的历史数据构建多元灰色模型如下:
    Figure PCTCN2021131990-appb-100001
    其中,τ表示延迟时间,r表示历史时序数,t表示时序数;t=1,2,…r表示时序数,
    Figure PCTCN2021131990-appb-100002
    为X i (0)的一次累加结果,
    Figure PCTCN2021131990-appb-100003
    为Y G (0)的一次累加结果,Y G (0)为多元灰色模型的输出预测序列;a和u均为模型参数,a表示控制系数,u表示灰色作用量;
    S12:求解多元灰色模型的响应函数,所述响应函数的计算公式为:
    Figure PCTCN2021131990-appb-100004
    其中,rf表示待预测的时序数,
    Figure PCTCN2021131990-appb-100005
    S13:对所述响应函数中的卷积积分进行离散化处理,得到运算结果之后进行一阶累减,得到多元灰色模型的输出预测序列如下:
    Y G (0)(τ+t)=Y G (1)(τ+t)-Y G (1)(τ+t-1);
    S14:计算多元灰色模型的输出预测序列与实际的电子固废产生量之间的预测误差如下:
    σ (0)=Y (0)-Y G (0)
    其中,Y G (0)={Y G (0)(τ+1),Y G (0)(τ+2),…,Y G (0)(τ+r)},
    σ (0)={σ (0)(τ+1),σ (0)(τ+2),…,σ (0)(τ+r)};
    S15:利用智能学习方法对多元灰色模型的预测误差进行智能补偿,得到误差补偿序列;
    S16:将误差补偿序列与多元灰色模型的输出预测序列相加,得到最终的电子固废产生量预测结果如下:
    Figure PCTCN2021131990-appb-100006
    其中,
    Figure PCTCN2021131990-appb-100007
    表示最终的电子固废产生量智能预测结果,
    Figure PCTCN2021131990-appb-100008
  3. 根据权利要求1所述的电子固废回收全流程智能解析方法,其特征在于:基于电子固废产生量预测值构建回收网点选址模型的方法包括:
    对所述电子固废产生量预测值进行归一化处理;
    基于归一化处理后的电子固废产生量预测值确定固废产生集中点;
    以最小化经济成本和最大化固废产生集中点覆盖率为目标构建回收网点选址模型。
  4. 根据权利要求1所述的电子固废回收全流程智能解析方法,其特征在于:根据所述回收网点位置分布构建分拣中心选址模型的方法包括:
    将回收网点和分拣中心作为上层,拆解企业作为下层,基于上下两层的相互联系和制约构建分拣中心选址模型。
  5. 根据权利要求1所述的电子固废回收全流程智能解析方法,其特征在于:基于所述回收网点位置分布构建回收人员调度模型的方法包括:
    将所述回收网点位置分布作为输入,回收人员调度方案作为输出构建回收人员调度模型。
  6. 根据权利要求1所述的电子固废回收全流程智能解析方法,其特征在于:基于所述回收网点位置分布和所述分拣中心位置分布构建回收车辆调度模型包括:
    对所述回收网点位置分布和所述分拣中心位置分布进行归一化处理;
    将归一化处理后的回收网点位置分布和所述分拣中心位置分布作为输入,回收车辆调度方案作为输出构建回收车辆调度模型。
  7. 一种电子固废回收全流程智能解析系统,其特征在于,包括:
    数据获取模块,所述数据获取模块用于获取待预测地区的电子固废产生量以及影响固废产生量的历史数据,对所述历史数据进行归一化处理,利用归一化处理后的历史数据构建多元灰色模型,利用智能学习算法对多元灰色模型的预测值进行误差补偿,获得待预测地区的电子固废产生量预测值;
    位置分布规划模块,所述位置分布规划模块用于基于电子固废产生量预测值构建回收网点选址模型,利用智能算法对回收网点选址模型进行求解,获得回收网点位置分布,并根据所述回收网点位置分布构建分拣中心选址模型,利用智能算法对分拣中心选址模型进行求解,获得分拣中心位置分布;
    智能调度模块,所述智能调度模块用于基于所述回收网点位置分布构建回收人员调度模型,基于所述回收网点位置分布和所述分拣中心位置分布构建回收车辆调度模型,基于所述分拣中心位置分布构建长期供货模型及短期调货模型,利用智能算法分别对回收人员调度模型、回收车辆调度模型、长期供货模型及短期调货模型进行求解,获得回收人员调度方案、回收车辆调度方案、长期供货方案及短期调货方案;
    可视化模块,所述可视化模块用于对所述回收人员调度方案、回收车辆调度方案、长期供货方案及短期调货方案进行可视化处理。
  8. 根据权利要求7所述的电子固废回收全流程智能解析系统,其特征在于:所述数据获取模块包括电子固废预测子模块,所述电子固废预测子模块用于利用归一化处理后的历史数据构建多元灰色模型,利用智能学习算法对多元灰色模型的预测值进行误差补偿,获得待预测地区的电子固废产生量预测值,包括:
    S11:利用归一化处理后的历史数据构建多元灰色模型如下:
    Figure PCTCN2021131990-appb-100009
    其中,τ表示延迟时间,r表示历史时序数,t表示时序数;t=1,2,…r表示时序数,
    Figure PCTCN2021131990-appb-100010
    为X i (0)的一次累加结果,
    Figure PCTCN2021131990-appb-100011
    为Y G (0)的一次累加结果,Y G (0)为多元灰色模型的输出预测序列;a和u均为模型参数,a表示控制系数,u表示灰色作用量;
    S12:求解多元灰色模型的响应函数,所述响应函数的计算公式为:
    Figure PCTCN2021131990-appb-100012
    其中,rf表示待预测的时序数,
    Figure PCTCN2021131990-appb-100013
    S13:对所述响应函数中的卷积积分进行离散化处理,得到运算结果之后进行一阶累减,得到多元灰色模型的输出预测序列如下:
    Y G (0)(τ+t)=Y G (1)(τ+t)-Y G (1)(τ+t-1);
    S14:计算多元灰色模型的输出预测序列与实际的电子固废产生量之间的预测误差如下:
    σ (0)=Y (0)-Y G (0)
    其中,Y G (0)={Y G (0)(τ+1),Y G (0)(τ+2),…,Y G (0)(τ+r)},
    σ (0)={σ (0)(τ+1),σ (0)(τ+2),…,σ (0)(τ+r)};
    S15:利用智能学习方法对多元灰色模型的预测误差进行智能补偿,得到误差补偿序列;
    S16:将误差补偿序列与多元灰色模型的输出预测序列相加,得到最终的电子固废产生量预测结果如下:
    Figure PCTCN2021131990-appb-100014
    其中,
    Figure PCTCN2021131990-appb-100015
    表示最终的电子固废产生量智能预测结果,
    Figure PCTCN2021131990-appb-100016
  9. 根据权利要求7所述的电子固废回收全流程智能解析系统,其特征在于:所述位置分布规划模块包括回收网点构建子模块,所述回收网点构建子模块用于基于电子固废产生量预测值构建回收网点选址模型,包括:
    对所述电子固废产生量预测值进行归一化处理;
    基于归一化处理后的电子固废产生量预测值确定固废产生集中点;
    以最小化经济成本和最大化固废产生集中点覆盖率为目标构建回收网点 选址模型。
  10. 根据权利要求7所述的电子固废回收全流程智能解析系统,其特征在于:所述位置分布规划模块包括分拣中心构建子模块,所述分拣中心构建子模块用于根据所述回收网点位置分布构建分拣中心选址模型,包括:
    将回收网点和分拣中心作为上层,拆解企业作为下层,基于上下两层的相互联系和制约构建分拣中心选址模型。
PCT/CN2021/131990 2021-11-12 2021-11-22 电子固废回收全流程智能解析方法及系统 WO2023082315A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202111339305.1 2021-11-12
CN202111339305.1A CN114077964B (zh) 2021-11-12 2021-11-12 电子固废回收全流程智能解析方法及系统

Publications (1)

Publication Number Publication Date
WO2023082315A1 true WO2023082315A1 (zh) 2023-05-19

Family

ID=80283788

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/131990 WO2023082315A1 (zh) 2021-11-12 2021-11-22 电子固废回收全流程智能解析方法及系统

Country Status (2)

Country Link
CN (1) CN114077964B (zh)
WO (1) WO2023082315A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116720623A (zh) * 2023-06-14 2023-09-08 西南交通大学 一种涉及站内工具更换的直线布局拆解方案设计方法

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105913213A (zh) * 2016-06-08 2016-08-31 沈阳工业大学 仓储集货模式下逆向物流回收车辆调度方法
CN111222652A (zh) * 2020-01-18 2020-06-02 江南大学 一种废弃电器电子产品回收流程中的智能匹配方法
CN111401655A (zh) * 2020-03-26 2020-07-10 上海第二工业大学 一种医疗废弃物回收设施选址的决策方法
CN113077099A (zh) * 2021-04-16 2021-07-06 国网安徽省电力有限公司 电网企业危险废物集中存贮中心选址和路径规划的方法

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20130075468A1 (en) * 2010-10-05 2013-03-28 Ever Green Environmental Corporation System and method for recycling of articles

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105913213A (zh) * 2016-06-08 2016-08-31 沈阳工业大学 仓储集货模式下逆向物流回收车辆调度方法
CN111222652A (zh) * 2020-01-18 2020-06-02 江南大学 一种废弃电器电子产品回收流程中的智能匹配方法
CN111401655A (zh) * 2020-03-26 2020-07-10 上海第二工业大学 一种医疗废弃物回收设施选址的决策方法
CN113077099A (zh) * 2021-04-16 2021-07-06 国网安徽省电力有限公司 电网企业危险废物集中存贮中心选址和路径规划的方法

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
"Master's Thesis", 1 April 2015, GUILIN UNIVERSITY OF ELECTRONIC TECHNOLOGY, CN, article ZHONG, HAILING: "Reverse Logistics Network Planning Study of Electronic Waste under Quality Uncertainty", pages: 1 - 67, XP009546313 *
SUN XIAOAN, XIAOLI LUAN, FEI LIU: "Electronic Solid Waste Prediction Based on Intelligent Optimization Grey Mode", JOURNAL OF SYSTEM SIMULATION, vol. 34, no. 3, 1 March 2022 (2022-03-01), pages 536 - 542, XP093066411, DOI: 10.16182/j.issn1004731x.joss.20-0833 *
曾庆成 等 (ZENG, QINGCHENG ET AL.): "配送中心选址与车辆路径一体优化模型与算法 (Optimization Model and Algorithm of Coordinated Distribution Center Location and Vehicle Routing)", vol. 33, no. 2, 15 April 2009 (2009-04-15), XP009546314 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116720623A (zh) * 2023-06-14 2023-09-08 西南交通大学 一种涉及站内工具更换的直线布局拆解方案设计方法

Also Published As

Publication number Publication date
CN114077964A (zh) 2022-02-22
CN114077964B (zh) 2022-07-15

Similar Documents

Publication Publication Date Title
Wang et al. Two-echelon location-routing optimization with time windows based on customer clustering
Mohtashami et al. A green closed loop supply chain design using queuing system for reducing environmental impact and energy consumption
Ardjmand et al. Applying genetic algorithm to a new bi-objective stochastic model for transportation, location, and allocation of hazardous materials
Zhalechian et al. A self-adaptive evolutionary algorithm for a fuzzy multi-objective hub location problem: An integration of responsiveness and social responsibility
Pamucar et al. Vehicle route selection with an adaptive neuro fuzzy inference system in uncertainty conditions
Mousavi et al. A new decision model for cross-docking center location in logistics networks under interval-valued intuitionistic fuzzy uncertainty
CN111401808A (zh) 一种基于混合模型的物资协议库存需求预测方法
Abdolazimi et al. Designing a new mathematical model based on ABC analysis for inventory control problem: A real case study
CN112884404A (zh) 一种智能供应链库存中转优化及异动预警系统
Rabbouch et al. Efficient implementation of the genetic algorithm to solve rich vehicle routing problems
Foroozesh et al. A hybrid decision-making method using robust programming and interval-valued fuzzy sets for sustainable-resilient supply chain network design considering circular economy and technology levels
Wang et al. Bi-objective green supply chain network design under disruption risk through an extended NSGA-II algorithm
Saha et al. A decision support system for real-time order management in a heterogeneous production environment
Rostamzadeh et al. Hybrid GA for material routing optimization in supply chain
Gitinavard et al. A bi-objective multi-echelon supply chain model with Pareto optimal points evaluation for perishable products under uncertainty
WO2023082315A1 (zh) 电子固废回收全流程智能解析方法及系统
Jafarian et al. The inventory-routing problem subject to vehicle failure
Mardanya et al. Solving bi-level multi-objective transportation problem under fuzziness
Cai et al. A survey of dynamic pickup and delivery problems
Chu et al. A Physarum-inspired algorithm for logistics optimization: From the perspective of effective distance
Jiang et al. Two-sided matching decision-making model for complex product system based on life-cycle sustainability assessment
Lee et al. Cloud-based cyber-physical logistics system with nested MAX-MIN ant algorithm for e-commerce logistics
Nozari et al. A Comprehensive Strategic-Tactical Multi-Objective Sustainable Supply Chain Model with Human Resources Considerations
Dotoli et al. Integrated network design of agile resource-efficient supply chains under uncertainty
Jun et al. Evolutionary neural network for learning of scalable heuristics for pickup and delivery problems with time windows

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21963759

Country of ref document: EP

Kind code of ref document: A1