Acquisition method for domain rule knowledge of industrial process

An acquisition method for domain rule knowledge of an industrial process, comprising:

• establishing a domain rule base, establishing a semantic knowledge base, and combining the domain rule base and the semantic knowledge base so as to realize an augmented update of a domain rule knowledge base;
• wherein steps of establishing the domain rule base comprise:
• performing representation and extraction of domain knowledge by describing domain knowledge of the industrial process by using weighted first-order logic rules so as to form a training sample set of the first-order logic rules;
• and
• performing construction of the domain rule base by performing a weight learning by applying probability soft logic and the training sample set of the first-order logic rules so as to realize weight to non-weighted rules;
• wherein steps of establishing the semantic knowledge base comprise:
• performing rule learning through a machine learning algorithm to obtain a first-order logic rule on a change in optimization decision-making semantics when multi-source data semantic information changes;
• wherein a specific form of the weighted first-order logic rules is that if an N-ary predicate P represents that a property, state, or attribute of a certain industrial process target D is DS under a condition of C1, . . . , Cn, P(D, C1, . . . , Cn, Ds) is an atom;
• a relationship between a certain optimization decision-making problem and a series of causes of the optimization decision-making problem expressed by the weighted first-order logic rules is as follows:
• ∀D1,D2, . . . ,Dl,R·P1(D1, . . . )∧P2(D2, . . . )∧ . . . ∧Pl(Dl, . . . )⇒PR(Q,R):λ,
• and the relationship indicates that states of industrial process targets D1, D2, . . . Dl in a certain condition cause the optimization decision-making problem Q resulting in R;
• wherein P1, P2, . . . , P1, PR are predicates;
• D1, D2, . . . Dl, R are variables;
• l represents a total number of predicate logical representations on the industrial process targets in one rule;
• and λ represents a weight;
• a specific form of semantic knowledge is k=(H1, H2, . . . , HN):{A1, A2, . . . , AN}:
• R, wherein a specific from of R is {R1, R2, . . . , RN};
• HN is an n-th compact characteristic of an n-th multi-source data XN;
• Ai is semantic information of an n-th multi-source data XN;
• in a case that different multi-source data corresponds to different semantic information changes, a variation law of the optimization decision-making semantics RN is implied;
• by mining a variation law between the semantic information and the optimization decision-making semantics of the different multi-source data from the semantic knowledge base, the first-order logic rule is extracted from the semantic information of the multi-source data;
• a specific step of performing the rule learning through the machine learning algorithm so as to obtain the first-order logic rule on the change in the optimization decision-making semantics when the multi-source data semantic information changes is:
• defining a logic predicate Pi(Hi, cAiij) to represent that characteristic semantics corresponding to an i-th compact characteristic Hi of the multi-source data is cAiij such that the first-order logic rule on the change in the optimization decision-making semantics when the multi-source data semantic information changes is obtained by performing the rule learning through the machine learning algorithm, and the first-order logic rule is represented by:
• P1(H1,cAiij)∧ . . . ∧PN(HN,cANNj)⇒PR(cRs):{circumflex over (λ)}
• the first-order logic rule indicates that a result of the optimization decision-making problem after the compact characteristics H1, H2, . . . , HN of the multi-source data being subject to the rule learning by using the machine learning algorithm is cRs,
• wherein PR(cRs) represents a logic predicate of the optimization decision-making semantics, and {circumflex over (λ)} represents a corresponding weight.