A project management system and implementation method based on industrial cloud platform

Abstract

The present invention provides a project management system based on an industrial cloud platform and an implementation method, comprising: a cloud industrial platform, users, and suppliers; wherein the cloud industrial platform comprises a database module for collecting user input data, an intelligent module for data processing, optimization, and generation of feasible solutions, and a decision support module for solution evaluation and comparison; the cloud platform collects user manufacturing service requests and supplier quotations, processes them through an intelligent decision-making algorithm, and finds the best manufacturing solution according to user needs; the intelligent decision-making algorithm generates solutions based on a nested algorithm, helping suppliers and users to minimize time and cost according to specific solution feature factors, and has good stability and system performance.

Description

Technical Field

[0001] The present invention relates to the industrial field, and more specifically, to a project management system and implementation method based on an industrial cloud platform. Background Art

[0002] Modern industry has a rapid development on a global scale, but user needs vary greatly, which has a great impact on resource efficiency. The industrial cloud platform can effectively promote the efficient use of industrial resources and provide on-demand manufacturing services to different users based on personalized services. At the industrial network level, through the comprehensive sharing and on-demand use of distributed computing, software, digital and resources on the cloud platform, an intelligent industrial network with higher resource efficiency, higher productivity and utilization can be created. Through the industrial cloud platform, users can access smart machines, production systems, and large amounts of data from different sources everywhere. However, there are still many technical issues to achieve efficient operation of the industrial cloud platform, including the development of a cloud-based platform, providing suitable interfaces for production systems / users, service-oriented automation utility delivery, and the processing of large amounts of data from different sources. Summary of the invention

[0003] In order to solve the above technical problems, the present invention proposes a project management system and implementation method based on an industrial cloud platform to help suppliers and users minimize time and cost according to specific solution feature factors, with good stability and system performance.

[0004] The present invention proposes a project management system based on an industrial cloud platform, comprising: a cloud industrial platform, users, and suppliers; wherein the cloud industrial platform comprises a database module for collecting user input data, an intelligent module for data processing, optimization, and generation of feasible solutions, and a decision support module for solution evaluation and comparison;

[0005] The cloud platform collects users’ manufacturing service requests and suppliers’ quotations, processes them through intelligent decision-making algorithms, and finds the best manufacturing solution based on user needs.

[0006] The intelligent decision algorithm generates solutions based on the nested algorithm. Each solution consists of multiple nested configurations, and each nested configuration contains the following parameters: processing time, energy consumption, and distance cost required to achieve pairing.

[0007] In the database module of the project management system of the industrial cloud platform, the size of the database is constantly expanding for practical application. Since these documents are written in everyday language with different expressions, how to quickly and accurately retrieve information from the database is still a huge challenge. In order to improve the efficiency and performance of retrieval, the vector space model method is adopted.

[0008] The database module consists of three main parts: processing part; query operation part; retrieval application part.

[0009] First, the processing part automatically extracts text information from the target collection and processes the collected text information through a defined sequence of operations, namely, tokenization, converting all words to lowercase, lemmatization, and removing stop words to build a content corpus.

[0010] Secondly, the query manipulation module works by reading and processing the given query. The processed query is pre-scanned to match its expansion with relevant words in the pre-defined risk-related vocabulary. The terms that are not found in the pre-defined risk-related vocabulary are expanded by using synonyms. The system then scans the terms in the original query and the expanded query, removing the terms that are not present in the content corpus.

[0011] Third, the retrieval application module combines the query and the corpus to calculate the query-document similarity.

[0012] The vector space model is used to represent text documents as identifier vectors and assign non-binary weights to the index terms in the query and document to calculate the similarity between each document and the query. j It can be represented as a t-dimensional vector. For the vector model, t is the total number of index items, and each dimension corresponds to a separate index item. The element w in each vector i,j is the term document pair (k i ,d j ) and w i,j ≥0. Wherein, as shown in the following formula:

[0013]

[0014]

[0015] w i,j The inverse frequency term is called the document frequency weight. If the weight vector model document d j yes The weight of a document can be quantified as shown below:

[0016]

[0017] where f i,j is the index item k in the document i The frequency of N is the total number of documents in the document set, n i is the index item k i The number of documents.

[0018] By using the vector space model and the inverse frequency term document frequency weight, document d jThe similarity between q and query q is quantified as follows:

[0019]

[0020] and are the norms of the document and query vectors, and is the inner product of the document and query vector.

[0021] Considering the extended query q e After the influence of , the improved calculation formula of the similarity between the query and the related cases is as follows:

[0022] score=sim(d j ,q o )+λ×sim(d j ,q e )

[0023] Where λ is q e Effect coefficient, 0<λ<1, take λ=0.7.

[0024] For any document d j The similarity between q and query q is calculated as follows:

[0025]

[0026] Since q0 and q e The overall similarity ranges from 0 to 1.7.

[0027] The database module of the industrial cloud platform first extracts text content information from the data set and processes it to generate a content corpus. After the user enters a query, the system begins to read and process the query, expands the original query using predefined related vocabulary, and filters out query terms that do not exist in the content corpus. Finally, the original query, expanded query, and content corpus are combined to calculate the query document similarity, and the top 10 similar cases are returned to the user.

[0028] The processing time required to achieve pairing is as follows:

[0029]

[0030] Among them, T p is the time required to prepare the data, T bja is the time required to build the job assembly, T s is the time required for setup, T BK is the build time of the configured build job, T R is the time machine and T required to remove the product from MCis the time required to replace the product.

[0031] The energy consumption is estimated as follows:

[0032]

[0033] In the formula, E uc is the supplier’s unit energy cost, which depends on the supplier’s location, and P is the maximum nominal power. BK Consider the time required for warm-up, cool-down, and cleaning.

[0034] The distance cost D of each pair is calculated based on the location of the user and supplier using a distance cost coefficient that takes into account the batch volume and weight. ij The distance cost is as follows:

[0035] D ij =δ D ||L sj -L ci ||.

[0036] Among them, Lsj and Lcj are the locations of suppliers and users, δ D is the distance cost factor.

[0037] Furthermore, in order to reduce cost and time, a particle swarm algorithm is used to measure and identify the impact of different resource utilization rates on decision performance. The main objective function based on duration and cost is shown in the following formula:

[0038]

[0039]

[0040] Resource utilization m Utilization duration Activity j Definition T m For a linear project schedule, resource utilization m is defined as N using the material cost of the duration activity j. j m , the daily cost rate is P j m , for resource utilization m and inactivity j, the supplier cost is A j m .

[0041] The random particle swarm of the particle swarm optimization algorithm initializes the system through an optimization search algorithm with update generation. The search space consists of m dimensions. The multidimensional vector represents the jth particle swarm as shown below:

[0042] Y j =(y j1 ,y j2 , ..., yjm );

[0043] The m-dimensional vector represents the velocity of the jth particle as shown below:

[0044] V j =(v j1 , v j2 , ..., v jm ).

[0045] According to the optimization problem, the fitness of each particle is evaluated. Based on the individual fitness, the optimal position of the jth particle is as follows:

[0046] P j =(p j1 , p j2 , ..., p jm ).

[0047] According to the global optimal position, the optimal individual position of the entire population is as follows:

[0048] G j =(g j1 , g j2 , ..., g jm )

[0049] The particle velocity and its position at each step are given by:

[0050] V j =w*V j +d1*s1*(Q j -Y j )+d2*s2*(GY j )

[0051] Y j =Y j +V j

[0052] From the equation above, the inertia weight is w, which governs the effect of the particle velocity. The random variables s1 and s2 are uniformly distributed in the range [0, 1]. The maximum step sizes controlled by the acceleration coefficients are d1 and d2.

[0053] Based on the existing speed, the new speed in the particle swarm optimization algorithm is calculated, and the best historical position and the current position distance are calculated. Each component value v is fixed with the particle search space [-vmax, vmax]. The particle will move to the current position according to the formula. If the stopping criteria are met, the process ends or repeats.

[0054] The local-global information sharing capability improves the global optimization performance based on the improved particle swarm algorithm. The particle velocity is randomly initialized according to the number of particles. Then, the fitness value of the particle is calculated. According to the fitness and position value of the particle, the global best particle is saved with the current single information, which improves the development and exploration speed of the global optimization process. The local-global information sharing is shown in the following formula:

[0055] V j =w*V j +d1*s1*(Q j -Y j )+d2*s2*(GY j )+d3*s3*|Q j -Y j |

[0056] φ3=d3*s3*|Q j -Y j |

[0057] Minimize project cost and time using particle counting algorithms.

[0058] The project management system and implementation method based on the industrial cloud platform provided by the present invention realize a smart manufacturing network with higher resource efficiency, realize dynamic sharing of manufacturing services, and provide them on demand according to a service-oriented model, so as to dynamically realize convenient sharing of various distributed resources according to the actual needs of users. In this framework, different users can search and request the required services from the manufacturing cloud, and dynamically assemble them into a manufacturing service solution. The platform adopts a modular structure, which is convenient for effective processing and management. The database module is specifically used for acquiring and storing user instances. The intelligent computing module first evaluates the functional compatibility between user instances, performs nested operations to generate geometric configurations, and finally calculates multiple attributes to generate and describe solutions. A decision support module is developed to help suppliers and users select the best solution based on specific solution feature factors, such as processing time, energy consumption, and distance cost. In addition, based on the improved particle swarm optimization algorithm, project time and cost are objective functions, which can minimize time and cost, and have good stability and system performance. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in combination with the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0060] The present invention proposes a project management system based on an industrial cloud platform, comprising: a cloud industrial platform, users, and suppliers; wherein the cloud industrial platform comprises a database module for collecting user input data, an intelligent module for data processing, optimization, and generation of feasible solutions, and a decision support module for solution evaluation and comparison;

[0061] The cloud platform collects users’ manufacturing service requests and suppliers’ quotations, processes them through intelligent decision-making algorithms, and finds the best manufacturing solution based on user needs.

[0062] The intelligent decision algorithm generates solutions based on the nested algorithm. Each solution consists of multiple nested configurations, and each nested configuration contains the following parameters: processing time, energy consumption, and distance cost required to achieve pairing.

[0063] In the database module of the project management system of the industrial cloud platform, the size of the database is constantly expanding for practical application. Since these documents are written in everyday language with different expressions, how to quickly and accurately retrieve information from the database is still a huge challenge. In order to improve the efficiency and performance of retrieval, the vector space model method is adopted.

[0064] The database module consists of three main parts: processing part; query operation part; retrieval application part.

[0065] First, the processing part automatically extracts text information from the target collection and processes the collected text information through a defined sequence of operations, namely, tokenization, converting all words to lowercase, lemmatization, and removing stop words to build a content corpus.

[0066] Secondly, the query manipulation module works by reading and processing the given query. The processed query is pre-scanned to match its expansion with relevant words in the pre-defined risk-related vocabulary. The terms that are not found in the pre-defined risk-related vocabulary are expanded by using synonyms. The system then scans the terms in the original query and the expanded query, removing the terms that are not present in the content corpus.

[0067] Third, the retrieval application module combines the query and the corpus to calculate the query-document similarity.

[0068] The vector space model is used to represent text documents as identifier vectors and assign non-binary weights to the index terms in the query and document to calculate the similarity between each document and the query. j It can be represented as a t-dimensional vector. For the vector model, t is the total number of index items, and each dimension corresponds to a separate index item. The element w in each vector i,j is the term document pair (k i ,d j) and w i,j ≥0. Wherein, as shown in the following formula:

[0069]

[0070]

[0071] w i,j The inverse frequency term is called the document frequency weight. If the weight vector model document d j yes The weight of a document can be quantified as shown below:

[0072]

[0073] where f i,j is the index item k in the document i The frequency of N is the total number of documents in the document set, n i is the index item k i The number of documents.

[0074] By using the vector space model and the inverse frequency term document frequency weight, document d j The similarity between q and query q is quantified as follows:

[0075]

[0076] and are the norms of the document and query vectors, and is the inner product of the document and query vector.

[0077] Considering the extended query q e After the influence of , the improved calculation formula of the similarity between the query and the related cases is as follows:

[0078] score=sim(d j ,q o )+λ×sim(d j ,q e )

[0079] Where λ is q e Effect coefficient, 0<λ<1, take λ=0.7.

[0080] For any document d j The similarity between q and query q is calculated as follows:

[0081]

[0082] Since q0 and q eThe overall similarity ranges from 0 to 1.7.

[0083] The database module of the industrial cloud platform first extracts text content information from the data set and processes it to generate a content corpus. After the user enters a query, the system begins to read and process the query, expands the original query using predefined related vocabulary, and filters out query terms that do not exist in the content corpus. Finally, the original query, expanded query, and content corpus are combined to calculate the query document similarity, and the top 10 similar cases are returned to the user.

[0084] The processing time required to achieve pairing is as follows:

[0085]

[0086] Among them, T p is the time required to prepare the data, T bja is the time required to build the job assembly, T s is the time required for setup, T BK is the build time of the configured build job, T R is the time machine and T required to remove the product from MC is the time required to replace the product.

[0087] The energy consumption is estimated as follows:

[0088]

[0089] In the formula, E uc is the supplier’s unit energy cost, which depends on the supplier’s location, and P is the maximum nominal power. BK Consider the time required for warm-up, cool-down, and cleaning.

[0090] The distance cost D of each pair is calculated based on the location of the user and supplier using a distance cost coefficient that takes into account the batch volume and weight. ij The distance cost is as follows:

[0091] D ij =δ D ||L sj -L ci ||.

[0092] Among them, Lsj and Lcj are the locations of suppliers and users, δ D is the distance cost factor.

[0093] Furthermore, in order to reduce cost and time, a particle swarm algorithm is used to measure and identify the impact of different resource utilization rates on decision performance. The main objective function based on duration and cost is shown in the following formula:

[0094]

[0095]

[0096] Resource utilization m Utilization duration Activity j Definition T m For a linear project schedule, resource utilization m is defined as N using the material cost of the duration activity j. j m , the daily cost rate is P j m , for resource utilization m and inactivity j, the supplier cost is A j m .

[0097] The random particle swarm of the particle swarm optimization algorithm initializes the system through an optimization search algorithm with update generation. The search space consists of m dimensions. The multidimensional vector represents the jth particle swarm as shown below:

[0098] Y j =(y j1 ,y j2 , ..., y jm );

[0099] The m-dimensional vector represents the velocity of the jth particle as shown below:

[0100] V j =(v j1 , v j2 , ..., v jm ).

[0101] According to the optimization problem, the fitness of each particle is evaluated. Based on the individual fitness, the optimal position of the jth particle is as follows:

[0102] P j =(p j1 , p j2 , ..., p jm ).

[0103] According to the global optimal position, the optimal individual position of the entire population is as follows:

[0104] G j =(g j1 , g j2 , ..., g jm )

[0105] The particle velocity and its position at each step are given by:

[0106] V j =w*V j +d1*s1*(Q j -Y j )+d2*s2*(GY j )

[0107] Y j =Y j +V j

[0108] From the equation above, the inertia weight is w, which governs the effect of the particle velocity. The random variables s1 and s2 are uniformly distributed in the range [0, 1]. The maximum step sizes controlled by the acceleration coefficients are d1 and d2.

[0109] Based on the existing speed, the new speed in the particle swarm optimization algorithm is calculated, and the best historical position and the current position distance are calculated. Each component value v is fixed with the particle search space [-vmax, vmax]. The particle will move to the current position according to the formula. If the stopping criteria are met, the process ends or repeats.

[0110] The local-global information sharing capability improves the global optimization performance based on the improved particle swarm algorithm. The particle velocity is randomly initialized according to the number of particles. Then, the fitness value of the particle is calculated. According to the fitness and position value of the particle, the global best particle is saved with the current single information. The development and exploration speed of the global optimization process is improved. The local-global information sharing is shown in the following formula:

[0111] V j =w*V j +d1*s1*(Q j -Y j )+d2*s2*(GY j )+d3*s B *|Q j -Y j |

[0112] φ3=d3*s3*|Q j -Y j |

[0113] Minimize project cost and time using particle counting algorithms.

[0114] The project management system and implementation method based on the industrial cloud platform provided by the present invention realize a smart manufacturing network with higher resource efficiency, realize dynamic sharing of manufacturing services, and provide them on demand according to a service-oriented model, so as to dynamically realize convenient sharing of various distributed resources according to the actual needs of users. In this framework, different users can search and request the required services from the manufacturing cloud, and dynamically assemble them into a manufacturing service solution. The platform adopts a modular structure, which is convenient for effective processing and management. The database module is specifically used for acquiring and storing user instances. The intelligent computing module first evaluates the functional compatibility between user instances, performs nested operations to generate geometric configurations, and finally calculates multiple attributes to generate and describe solutions. A decision support module is developed to help suppliers and users select the best solution based on specific solution feature factors, such as processing time, energy consumption, and distance cost. In addition, based on the improved particle swarm optimization algorithm, project time and cost are objective functions, which can minimize time and cost, and have good stability and system performance.

[0115] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A project management system based on an industrial cloud platform, characterized in that: include: Cloud industrial platforms, users, and suppliers; The cloud industrial platform includes a database module for collecting user input data, an intelligent module for data processing, optimization and generation of feasible solutions, and a decision support module for solution evaluation and comparison; The cloud platform collects users’ manufacturing service requests and suppliers’ quotations, processes them through intelligent decision-making algorithms, and finds the best manufacturing solution based on user needs; The intelligent decision algorithm generates manufacturing solutions based on the nesting algorithm. Each manufacturing solution consists of multiple nested configurations. Each nested configuration contains the following parameters: processing time, energy consumption, and distance cost required to achieve pairing; The database module consists of three parts: processing part; query operation part; retrieval application part; First, the processing part automatically extracts text information from the target collection and processes the collected text information through a defined sequence of operations: tokenization, converting all words to lowercase, lemmatization, and removing stop words to build a content corpus; Secondly, the query operation part is performed by reading and processing the given query; the processed query is pre-scanned to match its expansion with related words in the predefined risk-related vocabulary; the terms that are not found in the predefined risk-related vocabulary are expanded by using synonyms; then, the system scans the terms in the original query and the expanded query, and removes the terms that do not exist in the content corpus; Third, the retrieval application part combines the query and the corpus to calculate the query-document similarity; The vector space model is used to represent text documents as identifier vectors and assign non-binary weights to the index terms in the query and document to calculate the similarity between each document and the query; the query q and document d j Represented as a t-dimensional vector; for the vector model, t is the total number of index items, and each dimension corresponds to a separate index item; the element w in each vector i,j is the term document pair (k i ,d j ) and w i,j ≥0; where, as shown in the following formula: w i,j It is called the inverse frequency term document frequency weight; if the weight vector model document d j yes The document weight is quantified as shown below: Where fi,j is the frequency of index term ki in the document, N is the total number of documents in the document set, and ni is the number of documents containing index term k. i The number of documents; By using the vector space model and the inverse frequency term document frequency weight, document d j The similarity between q0 and query q0 is quantified as follows: and are the norms of the document and query vectors, and is the inner product of the document and query vector; After considering the impact of the extended query qe, the improved formula for calculating the similarity between the query and related cases is as follows: score=sim(d j ,q o )+λ×sim(d j ,q e ) Where λ is the expanded query q e Effect coefficient, 0<λ<1, take λ=0.7; For any document d j The similarity between q and query q is calculated as follows: Since q0 and q e The overall similarity ranges from 0 to 1.7; The processing time required to achieve pairing is as follows: T=T p +k1*T bja +k2*T S +∑ K T BK +k3*T R +T MC Among them, T p is the time required to prepare the data, T bja is the time required to build the job assembly, T s is the time required for setup, T BK is the build time of the configured build job, T R is the time and T required to remove the product from MC is the time required to replace the product; The energy consumption is estimated as follows: In the formula, E uc is the supplier’s single energy cost, which depends on the supplier’s location, and P is the maximum nominal power; The distance cost Di is calculated for each pairing based on the location of the user and supplier, using a distance cost coefficient that takes into account the batch volume and weight. j ; The distance cost is as follows: D ij =δ D ||L sj -L ci ||; Among them, Lsj and Lcj are the locations of suppliers and users, δ D is the distance cost coefficient; In order to reduce cost and time, the particle swarm algorithm is used to measure and identify the impact of different resource utilization rates on decision performance. The objective function based on construction period and cost is shown as follows: Resource utilization m Utilization duration Activity j Definition T m ; For a linear project schedule, the resource utilization m is defined as N using the duration activity j on the material cost j m , the daily cost rate is P j m , for resource utilization m and inactivity j, the supplier cost is A j m ; The random particle swarm of the particle swarm optimization algorithm initializes the system through an optimization search algorithm with update generation. The search space consists of n dimensions; the multidimensional vector represents the lth particle swarm as shown in the following formula: AND l (and l1 ,and l2 ,…,and ln ) The n-dimensional vector represents the velocity of the lth particle as shown below: V l =(v l1 ,v l2 ,…,v ln ) According to the optimization problem, the fitness of each particle is evaluated. Based on the individual fitness, the optimal position of the lth particle is shown as follows: Q l =(q l1 ,q l2 ,…,q ln ) According to the global optimal position, the optimal individual position of the entire population is as follows: G l =(g l1 ,g l2 ,…,g ln ) The particle velocity and its position at each step are given by: V l =w*V l +d1*s1*(Q l -Y l )+d2*s2*(G l -Y l ) AND l =And l +V l From the above equations, the inertia weight is w, managing the effect of particle velocity; the random variables s1 and s2 are uniformly distributed in the range [0, 1]; the maximum step sizes controlled by the acceleration coefficients are d1 and d2; According to the existing speed, the new speed in the particle swarm optimization algorithm is calculated, and the distance between the best historical position and the current position is calculated; each component value v is fixed with the particle search space [-vmax, vmax]; the particle will move to the current position according to the formula; if the stopping criteria are met, the process ends or repeats; The local-global information sharing capability improves the global optimization performance based on the improved particle swarm algorithm; the particle velocity is randomly initialized according to the number of particles; then, the fitness value of the particle is calculated; according to the fitness and position value of the particle, the global best particle is saved with the current single information to improve the development and exploration speed of the global optimization process; local-global information sharing is shown in the following formula: <h2 style=";text-align:left;direction:ltr">V<h2 style=";text-align:left;direction:ltr"> l <h2 style=";text-align:left;direction:ltr"> (w*V)<h2 style=";text-align:left;direction:ltr"> l <h2 style=";text-align:left;direction:ltr"> +d1*s1*(Q<h2 style=";text-align:left;direction:ltr"> l <h2 style=";text-align:left;direction:ltr"> -Y<h2 style=";text-align:left;direction:ltr"> l <h2 style=";text-align:left;direction:ltr"> )+d2*s2*(G<h2 style=";text-align:left;direction:ltr"> l <h2 style=";text-align:left;direction:ltr"> -Y<h2 style=";text-align:left;direction:ltr"> l <h2 style=";text-align:left;direction:ltr"> )+d3*s3*|Q<h2 style=";text-align:left;direction:ltr"> l <h2 style=";text-align:left;direction:ltr"> -Y<h2 style=";text-align:left;direction:ltr"> l <h2 style=";text-align:left;direction:ltr"> | Minimize project cost and time using particle counting algorithms.

Images