CN103295064A - Method for formalizing activity diagrams based on clinical practice guidelines - Google Patents

Abstract

The invention discloses a method for formalizing activity diagrams based on clinical practice guidelines. According to the method for formalizing the activity diagrams based on the clinical practice guidelines, an actual domain problem is converted into a problem of searching sub-diagrams in the activity diagrams by formalizing the activity diagrams based on the clinical practice guidelines. The method for formalizing the activity diagrams based on the clinical practice guidelines comprises performing symbol definition on the activity diagrams and the sub-diagrams based on clinical practice guidelines through a graph theory method so as to accurately matching objects of study, which the method can be applied to; for specific objects of study, namely, the activity diagrams, offering representing methods of programming languages of answer sets; and offering rule sets required by reasoning so as to build a complete answer set program for the reasoning. The whole process is intuitive and natural and facilitates operation.

Description

A kind of activity diagram based on clinical practice guideline is carried out formal method

Technical field

The invention belongs to the representation of knowledge field of artificial intelligence, relate to a kind of to the method based on the activity diagram formization of clinical medicine guide.

Background technology

M.Gelfond and V.Lifschitz have created Stable model semantics (Stable Model Semantics) in 1988 for the nonmonotonic logic programming, two people expanded to the logical program of extracting that comprises strong negative (Strong Negation) and weak negative (Weak Negation) with this model semantics afterwards, and renamed the Stable model semantics as Answer Set semanteme.Owing to answer collection program (answer set program, answer the collection program) the outstanding performance of language aspect non-monotonic reasoning, the program design mode that it is used as a kind of problem solving has been applied to a plurality of fields, and obtain many concrete application achievements, comprising: model detects, diagnosis, biological information, zoology and linguistics etc.One is answered the collection program is the set of some propositions (statements), wherein these propositions be describe one in the field object and these relation between objects in this field.Answer the collection program language and also have very strong extensibility, the semanteme that the collection program is answered in expansion is being explored by many research groups, to continually strengthen the ability to express of answering the collection program and to improve it in the terseness that solves the specific area problem.For example the expansion language based on classics answer collection program has the answer collection program that has preference, has probabilistic answer collection program, the cognitive collection program of answering, and CR-Prolog, action language (action language, AL) etc.

Answer collection program is made up of the rule of following form:

l ₀：-l ₁，...，l _m，not l _m+1，...，not l _n.

Here for

l _iIt all is the literal in the first order logic program language.{ l ₀Be the head of rule, { l ₁..., l _mBe the positive rule body of rule, { l _l..., l _mIt is the negative rule body of rule.The The reasoning results of answering the collection program is called answers collection.The meaning of above-mentioned regular expression is, " if we believe l ₁..., l _m, and have no reason to believe l _M+1..., l _n, we believe l so ₀" " not " in the rule be called " negating i.e. failure ", it exist for us and express and have unusual default information and provide great convenience.

At present, existing answer collection PROGRAM REASONING IN TEMPORAL LOGIC machine has DLV, Clingo, Smodels etc.These inference machines have been applied in the application of many representation of knowledge reasonings.For the convenience of programming, these inference machines are also supported the answer collection program syntax of many expansions, and polymerization operator (aggregation operators) is exactly the wherein important operation of a class.For example, " course of each student registration can not be less than 3, and can not more than 6 " can represent with following rule very easily:

3{enroll(S，C)：course(C)}6：-student(S).

Wherein, S represents the arbitrary element of student in gathering, and C represents the arbitrary element in the course set.

Summary of the invention

Technical matters: the invention provides intuitively nature of a kind of process, be convenient to operation, when running into a plurality of clinical practice guideline and time spent, the activity diagram to based on the clinical medicine guide of being convenient to expand support carries out formal method.

Technical scheme: of the present invention activity diagram based on clinical practice guideline is carried out formal method, comprise the steps:

1) with the digraph that has label on the limit in the graph theory, to the symbol definition of carrying out based on the activity diagram of clinical practice guideline, the result of definition is (CN ∪ AN ∪ DN ₀∪ DN _x, E, l:E → L), wherein:

CN, AN, DN ₀, DN _x, L is mutually disjoint set,

CN={x}, x is called context node,

Set A N, DN ₀, DN _x, the element of L is called action node, logical OR decision node, repulsion or decision node and label,

The V=CN ∪ AN ∪ DN here ₀∪ DN _x, the element of set E (x y) is called a limit, is the input limit of node y, and the limit is not imported for context node in the output limit of node x, for any non-decision node output limit at the most can only be arranged,

L is the label function, is the partial function that label is arrived on the limit, namely on the not every limit label is arranged, for

If x is decision node, then ((x, y)) always exists l;

2) to be reflected in the activity diagram be a subgraph to the candidate therapeutic scheme of supposition in the clinical practice guideline, defines subgraph with graph theory method, and the result of definition is:

The context node of activity diagram G belongs to subgraph H,

Arbitrary node among the subgraph H all can reach from context node,

For the arbitrary node x among the subgraph H, must there be at least one node y, y is the leaf node in the activity diagram, and node x can reach to node y,

For any repulsion among the subgraph H or decision node x, has only an output limit in subgraph H;

3) with answering collection program representation activity diagram, be designated as the dynamic module ∏ that answers the collection program _Ag, concrete grammar is that the fact below adding in answering the collection program constitutes dynamic module ∏ _Ag:

Come correspondence to represent each context node ct among the activity diagram ag with the following fact respectively, wherein cNode is the logical predicate name:

cNode(ag，ct).；

Come correspondence to represent each action node action among the activity diagram ag with the following fact respectively, wherein aNode is the logical predicate name:

aNode(ag，action).；

Come correspondence to represent each logical OR decision node orDecision among the activity diagram ag with the following fact respectively, wherein oNode is the logical predicate name:

oNode(ag，orDecision).；

Come correspondence to represent each repulsion or decision node orDecision among the activity diagram ag with the following fact respectively, wherein xNode is the logical predicate name:

xNode(ag，xorDecision).；

Come correspondence to represent every limit among the activity diagram ag with the following fact respectively, wherein edge is the logical predicate name, and x, y are respectively nodename:

edge(ag，x，y).；

Respectively with the following fact come correspondence represent among the activity diagram ag each have the limit of label (x, y), wherein label is the logical predicate name, x, y are respectively nodename, l is tag name:

label(ag，x，y，l).；

4) in answering the collection program, add inference rule, be designated as the nucleus module ∏ that answers the collection program _Core, concrete grammar is to add following rule in answering the collection program, formation nucleus module ∏ _Core:

With the nodes X among the following rule definition activity diagram Ag, wherein decisionNode and node are the new logical predicate names of introducing:

decisionNode(Ag，X)：-xNode(Ag，X).

decisionNode(Ag，X)：-oNode(Ag，X).

node(Ag，X)：-cNode(Ag，X).；

node(Ag，X)：-aNode(Ag，X).

node(Ag，X)：-decisionNode(Ag，X).

With the leaf node lNode among the following rule definition activity diagram Ag (Ag, X) and non-leaf node nLNode (Ag, X):

nLNode(Ag，X)：-edge(Ag，X，Y)，node(Ag，X)，node(Ag，Y).，

lNode(Ag，X)：-node(Ag，X)，not nLNode(Ag，X).

Wherein Y is the variable of expression arbitrary node;

With the candidate limit in the following rule definition activity diagram, the candidate limit in the described activity diagram, i.e. limit among the subgraph H:

For each repulsion or decision node xNode (Ag, X), to its an output limit edge of more options (Ag, X, Y) as candidate limit candidateEdge (Ag, X, Y), namely

0{candidateEdge(Ag，X，Y)：edge(Ag，X，Y)}1：-xNode(Ag，X).，

For each logical OR decision node oNode (Ag, X), select it all output limit edge (Ag, X, random subset Y) as candidate limit candidateEdge (Ag, X, Y), namely

{candidateEdgee(Ag，X，Y)：edge(Ag，X，Y)}：-oNode(Ag，X).，

For each node except decision node, its output limit edge (Ag, X, Y) at random elect as candidate limit candidateEdge (Ag, X, Y), namely

0{candidateEdge(Ag，X，Y)：edge(Ag，X，Y)}1：-node(Ag，X)，not decisionNode(Ag，X).；

The node that comprises with following rule definition subgraph:

nodeInTreatment(Ag，X)：-candidateEdge(Ag，X，Y).，

nodeInTreatment(Ag，Y)：-candidateEdge(Ag，X，Y).

Ag wherein, X, Y are variablees, nodeInTreatment is the new logical predicate name of introducing;

With internodal accessibility in the following rule definition subgraph:

reachable(Ag，X，X)：-node(Ag，X).

reachable(Ag，X，Y)：-candidateEdge(Ag，X，Y).，

reachable(Ag，X，Y)：-reachable(Ag，X，Z)，reachable(Ag，Z，Y).

Ag wherein, X, Y are variablees, reachable is the new logical predicate name of introducing;

The constraint condition that satisfies with following rule definition subgraph:

In the subgraph context node Cn must be arranged, namely

：-cNode(Ag，Cn)，not nodeInTreatment(Ag，Cn).，

Arbitrary node X must can reach from context node Cn, namely

：-nodeInTreatment(Ag，X)，cNode(Ag，Cn)，not reachable(Ag，Cn，X).

，

Arbitrary node X must reach at least one leaf node, namely

reachLeaf(Ag，X)：-reachable(Ag，X，Y)，lNode(Ag，Y).，

：-nodeInTreatment(Ag，X)，not reachLeaf(Ag，X).

Ag wherein, X, Y are variablees, reachLeaf is the new logical predicate name of introducing;

For auxiliary regular of the action node definition among the subgraph H: to any action node in the activity diagram, if in subgraph, just satisfy actionInTreatment (D, X), namely

actionInTreatment(D，X)：-nodeInTreatment(Ag，X)，aNode(Ag，X)，ag(D，Ag).。

Beneficial effect: the present invention compared with prior art has the following advantages:

Formalization of the present invention is based on the method for the activity diagram of clinical practice guideline, and the result after the formalization is that computing machine is executable.Therefore can be applied in the automedica aid decision-making system based on the formal result of the present invention.At present a lot of research groups just are being devoted to the formalization of clinical practice guideline, but also do not have a kind of language to become unified adopted industrywide standard.Some is that unavailable computing machine is carried out in the widely used formalization method, the formalization to basic medical procedures algorithm mainly paid close attention in these language, also has some to be not suitable for formalization complexity, the guideline structure that multistep is rapid.

The present invention changes into actual field question the problem of search subgraph in the activity diagram by the activity diagram of formalization based on the clinical treatment guide.The inventive method is at first carried out symbol definition with graph theory method to activity diagram and subgraph based on clinical practice guideline, in order to accurately match the research object that can use this method; To specific research object, namely activity diagram provides the method for expressing of answering the collection program language then; Then provide the needed rule set of reasoning, thereby construct the answer collection program of complete Gong reasoning.Whole process is nature intuitively, is convenient to operation.The answer collection program that the present invention obtains is that computing machine is executable, and widely used inference machine has smodels at present, clingo, DLV etc.In the practical application, the answer collection program that obtains is sent into the appeal inference machine, the output result is all qualified subgraphs that comprise in the activity diagram.Because activity diagram itself is based on the clinical practice guideline of medical field, Ding Yi the subgraph therapeutic scheme in the corresponding clinical practice guideline then in the present invention, so the present invention can be used for automatic medical decision making system, comes medical assistance field staff to do decision-making.

The inventive method is based on the method for pure description logic, has stronger extensibility, suitable formalization complexity, the guideline structure that multistep is rapid, above-mentioned answer collection program is divided into dynamic module and two modules of nucleus module in addition, wherein nucleus module does not rely on specific activities figure, has reusable characteristics.In addition, owing to answer the non-monotonic nature of collection program itself, method disclosed by the invention has higher appearance and becomes ability, this means, at first, the frequent abnormal conditions that occur of expression medical field that can be very natural; Secondly, when running into a plurality of clinical practice guideline and time spent, method of the present invention is convenient to expansion support.

Description of drawings

Fig. 1 is for answering the process flow diagram of collection program formization based on the activity diagram of clinical practice guideline.

Fig. 2 is the activity diagram based on the clinical practice guideline of duodenal ulcer.

Fig. 3 is the activity diagram based on the clinical practice guideline of TIA.

Embodiment

Provide below these two activity diagrams are carried out formal instantiation, remember that respectively the answer collection program that finally obtains is ∏ (du), ∏ (tia).

1) with the digraph that has label on the limit in the graph theory, to the symbol definition of carrying out based on the activity diagram of clinical practice guideline, the result of definition is (CN ∪ AN ∪ DN ₀∪ DN _x, E, l:E → L), wherein:

CN, AN, DN ₀, DN _x, L is mutually disjoint set,

CN={x}, x is called context node,

Set A N, DN ₀, DN _x, the element of L is called action node, logical OR decision node, repulsion or decision node and mark is answered,

The V=CN ∪ AN ∪ DN here ₀∪ DN _x, the element of set E (x y) is called a limit, is the input limit of node y, and the limit is not imported for context node in the output limit of node x, for any non-decision node output limit at the most can only be arranged,

L is the label function, is the partial function that label is arrived on the limit, namely on the not every limit label is arranged, for

If x is decision node, then ((x, y)) always exists l;

2) to be reflected in the activity diagram be a subgraph to the candidate therapeutic scheme of supposition in the clinical practice guideline, defines subgraph with graph theory method, and the result of definition is:

The context node of activity diagram G belongs to subgraph H,

Arbitrary node among the subgraph H all can reach from context node,

For the arbitrary node x among the subgraph H, must there be at least one node y, y is the leaf node in the activity diagram, and node x can reach to node y,

For any repulsion among the subgraph H or decision node x, has only an output limit in subgraph H;

3) represent two activity diagrams respectively with answering the collection program, be designated as the dynamic module ∏ that answers the collection program _Ag(du) and ∏ _Ag(tia), concrete grammar is that the fact below adding in answering the collection program constitutes dynamic module ∏ _Ag(du) and ∏ _Ag(tia):

∏ _Ag(du) building process:

Come correspondence to represent that each context in the activity diagram 2 saves with the following fact respectively, note Fig. 2 is duAg, and the context node among Fig. 2 is du:

cNode(duAg，du).；

Respectively with the following fact come correspondence represent in the activity diagram 2 each the action node, the action node among Fig. 2 has 5, is designated as sa respectively, et, ppi, sc, rs:

aNode(duAg，sa).

aNode(duAg，et).

aNode(duAg，ppi).；

aNode(duAg，sc).

aNode(duAg，rs).

Come correspondence to represent each logical OR decision node in the activity diagram 2 not occur the logical OR node among Fig. 2, so do not add any fact with the following fact respectively;

Come correspondence to represent each repulsion or decision node in the activity diagram 2 with the following fact respectively, the repulsion among Fig. 2 or decision node have 2, are designated as hpyTest and ulcerHealed respectively:

xNode(duAg，hpyTest).

xNode(dyAg，ulcerHealed).；

Respectively with the following fact come correspondence represent in the activity diagram 2 every limit edge (ag, x, y)., wherein, x, y is respectively nodename, has 7 limits among Fig. 2, the node that connects defined in the above:

edge(duAg，du，sa).

edge(duAg，sa，hpyTest).

edge(duAg，et，ulcerHealed).

edge(duAg，ppi，ulcerHealed).；

edge(duAg，hpyTest，et).

edge(duAg，hpyTest，ppi).

edge(duAg，ulcerHealed，sc).

Come correspondence to represent label label (ag on each limit in the activity diagram 2 with the following fact respectively, x, y, l)., wherein, x, y are respectively nodename, and l is tag name, and having among Fig. 2 on 4 limits has label, tag name is designated as hpp respectively, hpn, uh and unh, related node defined in the above:

label(duAg，hpyTest，et，hpp).

label(duAg，hpyTest，ppi，hpn).

label(duAg，ulcerHealed，sc，uh).；

label(duAg ulcerHealed rs unh)

In like manner, ∏ _Ag(tia) building process is as follows:

Come correspondence to represent that each context in the activity diagram 3 saves with the following fact respectively, note Fig. 2 is tiaAg, and the context node among Fig. 3 is tia:

cNode(tiaAg，tia).；

Respectively with the following fact come correspondence represent in the activity diagram 3 each the action node, the action node among Fig. 3 has 6, is designated as ec respectively, a, ts, pcs, d, nc:

aNode(tiaAg，ec).

aNode(tiaAg，a).

aNode(tiaAg，ts).

aNode(tiaAg，pcs).；

aNode(tiaAg，d).

aNode(tiaAg，nc).

Come correspondence to represent each logical OR decision node in the activity diagram 3 not occur the logical OR node among Fig. 3, so do not add any fact with the following fact respectively;

Come correspondence to represent each repulsion or decision node in the activity diagram 3 with the following fact respectively, the repulsion among Fig. 3 or decision node have 4, are designated as hypCaemia respectively, fast, and neuSymResolved and riskStroke:

xNode(tiaAg，hypCaemia).

xNode(tiaAg，fast).

xNode(tiaAg，neuSymRe solved).；

xNode(tiaAg，riskStroke).

Respectively with the following fact come correspondence represent in the activity diagram 3 every limit edge (ag, x, y)., x wherein, y is respectively nodename, it has 12 limits among Fig. 3, the node that connects defined in the above:

edge(tiaAg，tia，hypCaemia).

edge(tiaAg，a，riskStroke).

edge(tiaAg，ts，nc).

edge(tiaAg，d，nc).

edge(tiaAg，hypCaemia，fast).

edge(tiaAg，hypCaemia，ec).

edge(tiaAg，fast，pcs).；

edge(taiAg，fast，neuSymRe solved).

edge(tiaAg，neuSymRe solved，a).

edge(tiaAg，neuSymRe solved，ts).

edge(tiaAg，tiskStroke，pcs).

edge(tiaAg，riskStroke，d).

Respectively with the following fact come correspondence represent on each limit in the activity diagram 3 label label (ag, x, y, l).,

Wherein, x, y are respectively nodename, and l is tag name, and having 8 limits among Fig. 3 has label, and tag name is designated as ha respectively, hp, and fn, fp, nsr, nsnr, rsn and rse, related node defined in the above:

label(tiaAg，hypCaemia，fast，ha).

label(tiaAg，hypCaemia，ec，hp).

label(tiaAg，fast，pcs，fn).

label(tiaAg，fast，neuSymRe solved，fp).；

label(tiaAg，neuSymRe solved，a，nsr).；

label(tiaAg，neuSymRe solved，ts，nsnr).

label(tiaAg，riskStroke，pcs，rsn).

label(tiaAg，riskStroke，d，rse).

4) in answering the collection program, add inference rule, be designated as the nucleus module ∏ that answers the collection program _Core, this module is reusable, therefore what activity diagram Fig. 2 and Fig. 3 are added is identical rule set.Concrete grammar is to generate following rule in answering the collection program, formation nucleus module ∏ _Core:

With the nodes X among the following rule definition activity diagram Ag:

decisionNode(Ag，X)：-xNode(Ag，X). (1)

decisionNode(Ag，X)：-oNode(Ag，X). (2)

node(Ag，X)：-cNode(Ag，X). (3)，

node(Ag，X)：-aNode(Ag，X). (4)

node(Ag，X)：-decisionNode(Ag，X). (5)

Wherein regular (1), (2) expressions " logical OR node and repulsion or node are referred to as decision node ",

Rule (3), (4), (5) expression " context node, action node and decision node are referred to as node ".

With the leaf node lNode among the following rule definition activity diagram Ag (Ag, X) and non-leaf node nLNode (Ag, X):

nLNode(Ag，X)：-edge(Ag，X，Y)，node(Ag，X)，node(Ag，Y). (6)

lNode(Ag，X)：-node(Ag，X)，not nLNode(Ag，X). (7)

，

Wherein Y is the variable of expression arbitrary node, rule (6) expression " for arbitrary node X among the activity diagram Ag; if exist nodes X to arrive the limit of node Y; then nodes X is not leaf node ", rule (7) expression " for arbitrary node X among the activity diagram Ag; be not leaf node if there is evidence to show, then think leaf node ";

With the candidate limit in the following rule definition activity diagram, the candidate limit in the described activity diagram, i.e. limit among the subgraph H:

For each repulsion or decision node xNode (Ag, X), to its an output limit edge of more options (Ag, X, Y) as candidate limit candidateEdge (Ag, X, Y), namely

0{candidateEdge(Ag，X，Y)：edge(Ag，X，Y)}1：-xNode(Ag，X). (8)

，

For each logical OR decision node oNode (Ag, X), select it all output limit edge (Ag, X, random subset Y) as candidate limit candidateEdge (Ag, X, Y), namely

{candidateEdge(Ag，X，Y)：edge(Ag，X，Y)}：-oNode(Ag，X). (9)，

For each node except decision node, its output limit edge (Ag, X, Y) at random elect as candidate limit candidateEdge (Ag, X, Y), namely

0{candidateEdge(Ag，X，Y)：edge(Ag，X，Y)}1：-node(Ag，X)，not decisionNode(Ag，X). (10)；

The node that comprises with following rule definition subgraph:

nodeInTreatment(Ag，X)：-candidateEdge(Ag，X，Y). (11)

nodeInTreatment(Ag，Y)：-candidateEdge(Ag，X，Y). (12)

Ag wherein, X, Y are variablees, rule (11), (12) expression " node that the candidate limit connects is all in subgraph ";

With internodal accessibility in the following rule definition subgraph:

reachable(Ag，X，X)：-node(Ag，X). (13)

reachable(Ag，X，Y)：-candidateEdge(Ag，X，Y). (14)

reachable(Ag，X，Y)：-reachable(Ag，X，Z)，reachable(Ag，Z，Y). (15)

，

Wherein rule (13) expression " arbitrary node X can reach to self ", rule (14) expression " can reach between the node that the candidate limit connects ", rule (15) expression " accessibility of node has transitivity, and even nodes X can reach to node Z; node Z can reach to node Y, and then nodes X can reach to node Y ";

The constraint condition that satisfies with following rule definition subgraph:

In the subgraph context node Cn must be arranged, namely

：-cNode(Ag，Cn)，not nodeInTreatment(Ag，Cn).，

Arbitrary node X must can reach from context node Cn, namely

：-nodeInTreatment(Ag，X)，cNode(Ag，Cn)，not reachable(Ag，Cn，X).

，

Arbitrary node X must reach at least one leaf node, namely

reachLeaf(Ag，X)：-reachable(Ag，X，Y)，lNode(Ag，Y).

：-nodeInTreatment(Ag，X)，not reachLeaf(Ag，X).；

For auxiliary regular of the action node definition among the subgraph H: to any action node in the activity diagram, if in subgraph, just satisfy actionInTreatment (D, X), namely

actionInTreatment(D，X)：-nodeInTreatment(Ag，X)，aNode(Ag，X)，ag(D，Ag).。

Claims (1)

1. one kind is carried out formal method to the activity diagram based on clinical practice guideline, it is characterized in that this method comprises the steps:

1) with the digraph that has label on the limit in the graph theory, to the symbol definition of carrying out based on the activity diagram of clinical practice guideline, the result of definition is (CN ∪ AN ∪ DN ₀∪ DN _x, E, l:E → L), wherein:

CN, AN, DN ₀, DN _x, L is mutually disjoint set,

CN={x}, x is called context node,

Set A N, DN ₀, DN _x, the element of L is called action node, logical OR decision node, repulsion or decision node and label,

The V=CN ∪ AN ∪ DN here ₀∪ DN _x, the element of set E (x y) is called a limit, is the input limit of node y, and the limit is not imported for context node in the output limit of node x, for any non-decision node output limit at the most can only be arranged,

L is the label function, is the partial function that label is arrived on the limit, namely on the not every limit label is arranged, for

If x is decision node, then ((x, y)) always exists l;

2) to be reflected in the activity diagram be a subgraph to the candidate therapeutic scheme of supposition in the clinical practice guideline, defines subgraph with graph theory method, and the result of definition is:

The context node of activity diagram G belongs to subgraph H,

Arbitrary node among the subgraph H all can reach from context node,

For the arbitrary node x among the subgraph H, must there be at least one node y, y is the leaf node in the activity diagram, and node x can reach to node y,

For any repulsion among the subgraph H or decision node x, has only an output limit in subgraph H.

3) with answering collection program representation activity diagram, be designated as the dynamic module ∏ that answers the collection program _Ag, concrete grammar is that the fact below adding in answering the collection program constitutes dynamic module ∏ _Ag:

Come correspondence to represent each context node ct among the activity diagram ag with the following fact respectively, wherein cNode is the logical predicate name:

cNode(ag,ct).；

Come correspondence to represent each action node action among the activity diagram ag with the following fact respectively, wherein aNode is the logical predicate name:

aNode(ag,action).；

Come correspondence to represent each logical OR decision node orDecision among the activity diagram ag with the following fact respectively, wherein oNode is the logical predicate name:

oNode(ag,orDecision).；

Come correspondence to represent each repulsion or decision node orDecision among the activity diagram ag with the following fact respectively, wherein xNode is the logical predicate name:

xNode(ag,xorDecision).；

Come correspondence to represent every limit among the activity diagram ag with the following fact respectively, wherein edge is the logical predicate name, and x, y are respectively nodename:

edge(ag,x,y).；

Respectively with the following fact come correspondence represent among the activity diagram ag each have the limit of label (x, y), wherein label is the logical predicate name, x, y are respectively nodename, l is tag name:

label(ag,x,y,l).；

4) in answering the collection program, add inference rule, be designated as the nucleus module ∏ that answers the collection program _Core, concrete grammar is to add following rule in answering the collection program, formation nucleus module ∏ _Core:

With the nodes X among the following rule definition activity diagram Ag, wherein decisionNode and node are the new logical predicate names of introducing:

decisionNode(Ag,X):-xNode(Ag,X).

decisionNode(Ag,X):-oNode(Ag,X).

node(Ag,X):-cNode(Ag,X).；

node(Ag,X):-aNode(Ag,X).

node(Ag,X):-decisionNode(Ag,X).

With the leaf node lNode among the following rule definition activity diagram Ag (Ag, X) and non-leaf node nLNode (Ag, X):

$\begin{array}{c}\mathrm{nLNode}(\mathrm{Ag},X):-\mathrm{edge}(\mathrm{Ag},X,Y),\mathrm{node}(\mathrm{Ag},X),\mathrm{node}(\mathrm{Ag},Y).\\ \mathrm{lNode}(\mathrm{Ag},X):-\mathrm{node}(\mathrm{Ag},X),\mathrm{not\; nLNode}(\mathrm{Ag},X).\end{array},$ Wherein Y is the variable of expression arbitrary node;

With the candidate limit in the following rule definition activity diagram, the candidate limit in the described activity diagram, i.e. limit among the subgraph H:

For each repulsion or decision node xNode (Ag, X), to its an output limit edge of more options (Ag, X, Y) as candidate limit candidateEdge (Ag, X, Y), namely

0{candidateEdge(Ag,X,Y):edge(Ag,X,Y)}1:-xNode(Ag,X).，

For each logical OR decision node oNode (Ag, X), select it all output limit edge (Ag, X, random subset Y) as candidate limit candidateEdge (Ag, X, Y), namely

{candidateEdge(Ag,X,Y):edge(Ag,X,Y)}:-oNode(Ag,X).，

For each node except decision node, its output limit edge (Ag, X, Y) at random elect as candidate limit candidateEdge (Ag, X, Y), namely

0{candidateEdge(Ag,X,Y):edge(Ag,X,Y)}1:-node(Ag,X),not decisionNode(Ag,X).

；

The node that comprises with following rule definition subgraph:

$\begin{array}{c}\mathrm{nodeInTreatment}(\mathrm{Ag},X):-\mathrm{candidateEdge}(\mathrm{Ag},X,Y).\\ \mathrm{nodeInTreatment}(\mathrm{Ag},Y);-\mathrm{candidateEdge}(\mathrm{Ag},X,Y).\end{array},$ Ag wherein, X, Y are variablees, nodeInTreatment is the new logical predicate name of introducing;

With internodal accessibility in the following rule definition subgraph:

reachable(Ag,X,X):-node(Ag,X).

Reachable (Ag, X, Y) :-candidateEdge (Ag, X, Y)., wherein

reachable(Ag,X,Y):-reachable(Ag,X,Z),reachable(Ag,Z,Y).

Ag, X, Y are variablees, reachable is the new logical predicate name of introducing;

The constraint condition that satisfies with following rule definition subgraph:

In the subgraph context node Cn must be arranged, namely

:-cNode(Ag,Cn),not nodeInTreatment(Ag,Cn).，

Arbitrary node X must can reach from context node Cn, namely

:-nodeInTreatment(Ag,X),cNode(Ag,Cn),not reachable(Ag,Cn,X).

，

Arbitrary node X must reach at least one leaf node, namely

$\begin{array}{c}\mathrm{reachLeaf}(\mathrm{Ag},X):-\mathrm{reachable}(\mathrm{Ag},X,Y),\mathrm{lNode}(\mathrm{Ag},Y).\\ :-\mathrm{nodeInTreatment}(\mathrm{Ag},X),\mathrm{not\; reachLeaf}(\mathrm{Ag},X).\end{array},$ Ag wherein, X, Y are variablees, reachLeaf is the new logical predicate name of introducing;

For auxiliary regular of the action node definition among the subgraph H: to any action node in the activity diagram,

If in subgraph, just satisfy actionInTreatment (D, X), namely

actionInTreatment(D,X):-nodeInTreatment(Ag,X),aNode(Ag,X),ag(D,Ag).。

Images