Database and expert systems applications

D Karagiannis (ed.) Database and Expert Syste1ns Applications Proceedings of the International Conference in Berlin, Federal Republic of Germany, 1991 Springer-Verlag Wi en GmbH Dr.-Ing D Karagiannis Forschungsinstitut fur anwendungsorientierte Wissensverarbeitung (FAW) Ulm, Federal Republic of Germany This work is subject ta copyright Ali rights are reserved, whether the whole or part of the material is concerned, specifically those of translation, reprinting, re-use of illustrations, broadcasting, reproduction by photacopying machines or similar means, and storage in data banks © 1991 by Springer-Verlag Wien Originally published by Springer-Verlag Wien New York in 1991 With 351 Figures ISBN 978-3-211-82301-9 ISBN 978-3-7091-7555-2 (eBook) DOI 10.1007/978-3-7091-7555-2 Preface The Database and Expert Systems Applications - DEXA - conferences are dedicated to providing an international forum for the presentation of applications in the database and expert systems field, for the exchange of ideas and experiences, and for defining requirements for the future systems in these fields After the very promising DEXA 90 in Vienna, Austria, we hope to have successfully established wjth this year's DEXA 91 a stage where scientists from diverse fields interested in application-oriented research can present and discuss their work This year there was a total of more than 250 submitted papers from 28 different countries, in all continents Only 98 of the papers could be accepted The collection of papers in these proceedings offers a cross-section of the issues facing the area of databases and expert systems, i.e., topics of basic research interest on one hand and questions occurring when developing applications on the other Major credit for the success of the conference goes to all of our colleagues who submitted papers for consideration and to those who have organized and chaired the panel sessions Many persons contributed numerous hours to organize this conference The names of most of them will appear on the following pages In particular we wish to thank the Organization Committee Chairmen Johann Gordesch, A Min Tjoa, and Roland Wagner, who also helped establishing the program Special thanks also go to Gabriella Wagner and Anke Ruckert Dimitris Karagiannis General Conference Chairman Contents Conference Committee xiii List of Referees xiv SESSION A Object Oriented Representation Ramos I., O Pastor, V Casado; Spain "00 and Active Formal Information System Specification" Papazoglou M.P.; Australia "Roles: A Methodology for Representing Multifaceted Objects" Villemin F.-Y., A Paoli, I Tourrilhes, M Le; France "The ALDous90 Project: Merging Object-Oriented Databases and Knowledge-Based Systems" Burneau J.C., O Thiery; France "Object Oriented Databases for Maintenance Expert Systems" SESSION B 13 19 Office Information Systems O'Neill P., A AI-Zobaidie; Great Britain "Student Admission: Expert and Database Systems" 24 Guardalben G.V.; Italy "Mapping Generalization and Object Sharing into Nested Relations: An ODA Implementation" Hawryszkiewycz LT.; Australia "An Object Modeling Method for Office Information Systems Design" 28 34 Nakano Y., H Tanigawa, Y Imai; Japan "An Interface/Application Builder with Transmission Control Facility for SGML Document Databases" 41 SESSION C Deductive Databases and Database Programming Languages Adiba M., C Lecluse, P Richard; France "Rationale and Design of Serendip, a Database Programming Language" 47 Eder J.; Austria "General Transitive Closures and Aggregate Functions" Taylor M.C., B.D Czejdo; USA "A Deductive Database System with Applications to Route Planning" Suzuki S., T Ibaraki, M Kishi; Japan "Using Relaxation Techniques to Evaluate Queries in Deductive Databases" SESSION A 54 60 67 Applications on Technology and Industry Eum D., T Minoura; South Korea/USA "Data-Structure Builder for VLSI/CAD Software" 73 Tatsiopoulos I.P., N.D Mekras; Greece "Knowledge Representation Structures for the Evaluation of Production Planning and Control Systems" 80 l3aumewerd-Ahlmann A., A.B Cremers, G Kriiger, J Leonhardt, L Pliimer, R Waschkowski; Germany '·AII Information System for the Mining Industry" 86 van Oorschot J., W Korremans, K Vos; Netherlands "Intelligent Network Management: A Network Management Support System Using AI Techniques" 92 VIII SESSION B Information Retrieval Hartel R., P Lauppert; Germany "Computer-Aided Production of Multilingual Historical Reference Works" 98 Keirn D.A., K.-C Kim, V Lum; USA "A Friendly and Intelligent Approach to Data Retrieval in a Multimedia DBMS" 102 Fuhr N.; Germany "An Information Retrieval View of Environmental Information Systems" 112 Wendlandt E.B., J R Driscoll; USA "Enhancing Text Retrieval Semantically" SESSION C 118 Business and Educational Applications Goldstein R.C., V.C Storey; USA liThe Commonsense Business Reasoner" 124 Cleary D., J Zeleznikow; Australia "Building Expert Databases: L-CATA - An Intelligent Logic Based Travel Assistant" 1W LeU C.; Austria "Using a Meta-Knowledge Method for Developing an Educational Knowledge-Based Application" 135 Kamel M.N., G.A Zolla; USA "TARPS: A Prototype Expert Database System for Training and Administration of Reserves Officer Placement" 142 SESSION A Knowledge Based Systems Najmann 0., K Eckstein; Germany "Constructing Minimal Knowledge Bases by Machine Learning" Garidis C., S Bottcher; Germany "EffiCIent Access to Large Prolog Knowledge Bases" 154 Lasala P.; Spain "The ARPO-2: An Expert System for Building Contracts" SESSION B 160 Graphical Interfaces Coenen F., T Bench-Capon; Great Britain "A Graphical Interactive Tool for KBS Maintenance" Auddino A., E Amiel, B Bhargava; Switzerland/France/USA "Experiences with SUPER, a Database Visual Environment" SESSION A 148 1~ In Implementation Aspects Kriegel H.-P., P Heep, S Heep, M Schiwietz, R Schneider; Germany "A Flexible and Extensible Index Manager for Spatial Database Systems" Kumar V., J Mumper; USA "KRISHNA - Concurrency Control Algorithms Based on Dynamic Attributes" 179 1~ Touir A.; France "The Design of an Efficient Data Structure for Manipulating Data in an Image Database System" 191 Duppel N.; Germany "High Performance Data Parallel Recursion" SESSION B 197 Object Orientation Collet C., E Brunei; France "A Form System for an Object-Oriented Database System" 203 IX Staes F., L Tarantino, B Verdonk, D Vermeir; The Netherlands/Italy/Belgium "Supporting User Interactions with OODB's: A Declarative Approach" 210 Wu X.; Germany "A Query Interface for an Object Management System" 216 Fotouhi F., T.G Lee, W.r Grosky; USA "An Indexing Model for Complex Object Hierarchies in Object-Oriented Databases" 221 Barros 0., G Pavez; Chile "An Object-Oriented Executable Requirements Specification Language" SESSION C 227 Foundation of Object Orientation Canos J H., O Pastor; Spain "Functional and Object-Oriented Specification of Information Systems" 233 Cellary W., G Jomier, T Koszlajda; Poland/France "Formal Model of an Object-Oriented Database with Versioned Objects and Schema" 239 Lou Y., Z.M Ozsoyoglu; USA "Meta Variables and Inheritance in an Object-Oriented Data Model" 245 Tanaka K., S Kobayashi, T Sakanoue; Japan "Uncertainty Management in Object-Oriented Database Systems" 251 Kemper A., G Moerkotte; Germany "A Framework for Strong Typing and Type Inference in (Persistent) Object Models" SESSION A 257 Multimedia Databases and Hypertext Rhiner M.; Switzerland "Principles of Knowledge Augmented Visual Databases" 264 Hara Y., A.M Keller, G Wiederhold; USA "Relationship Abstractions for an Effective Hypertext Design: Augmentation and Globalization" 270 Delfs H., K Winkelmann; Germany "TOROS-HYPER: A Tool for the Integration of Hyper Documents into Knowledge-Based Systems" 275 van der Spiegel P.L., J.Th.W Driessen, P.D Bruza, Th.P van der Weide; Netherlands "A Transaction Model for Hypertext" 281 SESSION B Applications in Science, Technology, and Industry Martin T.P., J.r Glasgow, M.P Feret, T Kelley; Canada "A Knowledge-Based System for Fault Diagnosis in Real-Time Engineering Applications" Breunig M., G Heyer, A Perk hoff, M Seewald; Germany "An Expert System to Support Mine Planning Operations" 287 293 Meyer R., G Schlageter; Germany "Knowledge Base Management Systems in the Field of High Energy Physics" 299 Desmarais M., A Leblanc; Canada "A Meteorological Database for Numerical and Non-numerical Processing" SESSION C 305 Heterogenous and Multidatabase Systems Rusinkiewicz M., B Czejdo, D.W Embley; USA "An Implementation Model for Multidatabase Queries" Yang J., M.P Papazoglou, L Marinos; Australia/Netherlands "Knowledge-Based Schema Analysis in a Multi-Database Framework" 309 315 x Vidyasankar K.; Canada "Atomicity of Global Transactions in Distributed Heterogeneous Database Systems" 321 Endrikat A., R Michalski; Germany "Application-Oriented Integration of Distributed Heterogeneous Knowledge Sources" 327 SESSION A Advanced Applications Palomar M., L Moreno, A Pascual; Spain "Semantic Interpretation of Natural Language in PROLOG: Logical Forms" 333 Shimada S., H Matsushima; Japan "Map Summarization Using Analogical Matching of Schema" 339 Kersten M.L.; Netherlands "Goblin: A DBPL Designed for Advanced Database Applications" SESSION B 345 Modelling Atzeni P., R Torlone; Italy "A Metamodel Approach for the Management of Multiple Models in CASE Tools" 350 Wang X.Y., N.J Fiddian, W.A Gray; Great Britain "The Development of a Knowledge-Based Database Transaction Design Assistant" 356 Grefen P.W.P.J., P.M.G Apers; Netherlands "Integrity Constraint Enforcement Through Transaction Modification" 362 Carlson A.B., B.G Lundberg; Sweden "Environmental Information Organization" SESSION C 368 Database Aspects Andlinger P., C Gierlinger, G Quirchmayr; Austria "Making C++ Object Persistent by Using a Standard Relational Database System" Pernul G., K Karlapalem, S.B Navathe; USA "Relational Database Organization Based on Views and Fragments" 374 380 Morzy T.; Poland "A Dynamic Overwrite Protocol for Multiversion Concurrency Control Algorithms" SESSION A 387 Querying Aspects Hiilsmann K.; Germany "A Temporal-Logic-Based Query Language for Querying Database Histories" ~3 Pirri F., C Pizzuti; Italy "Querying Incomplete Knowledge Bases with Abduction" 400 Ozsoyoglu G., K Du, A Tjahjana, W-C Hou, D.Y Rowland; USA "On Estimating COUNT, SUM, and AVERAGE Relational Algebra Queries" 406 Schneider M., C Trepied; France "Recursion with the Graphical Query Language CANDID" Ferri F., P Grifoni., L Meo-Evoli, F.L Ricci; Italy "ADAMS: An Aggregate Data Management System with Multiple Interaction Techniques" SESSION B 413 419 Legal Systems Galindo F.; Spain "Juridical Principles for Juridical Applications The DERINFO Methodology" 425 XI Baaz M., G Quirchmayr; Austria "A Formal Model for the Support of Analogical Reasoning in Legal Expert Systems" 431 Bratley P., D Poulin, J Savoy; Canada "The Effect of Change on Legal Applications" 436 Basili C., N Palazzolo, A.M Tammaro; Italy "An Expert Thesaurus for Ancient Law" SESSION C 442 Hypermedia Systems Frei H.P., P Schauble; Switzerland "Designing a Hypermedia Information System" 449 Fuller M., A Kent, R Sacks-Davis, J Thom, R Wilkinson, J Zobel; Australia "Querying in a Large Hyperbase" 455 Wang B., P Hitchcock; Great Britain "InterSect: A General Purpose Hypertext System Based on an Object Oriented Database" 459 Eherer S., M Jarke; Germany "Knowledge Base Support for Hypermedia Co-Authoring" 465 Rugelj J.; Slovenia "Computer Supported Multimedia Environment for Collaboration" SESSION A U1 Practice Systems Zaring P.A., H.G Vogt; Sweden "Expert Systems Development and Management, A Case Study" 476 Tanaka H.; Japan "Protein Function Database as a Deductive and Object-Oriented Database" 481 Steyer F.; Germany "System Architecture and Specification of a Fast BOM Object Processor Using a Standard Relational Database Management System and a Main Memory Cache" 487 Scheer A.-W., S Spang; Germany "Data Modelling for International Organizations" 491 Kvalem J., R.-E Grini, K Haugset; Norway '1SACS - An Integrated Surveillance and Control System" 497 Scholz J., P Levi; Germany "Constraint-Guided Scheduling for Satellite Data Processing and A.rchiving" SESSION B 501 User Interfaces Risch T., G Wiederhold; USA "Building Adaptive Applications Using Active Mediator" 508 Signore 0., R Aulisi, V Ceccanti; Italy "Hypertext for Hypertext: A Figured Thesaurus" 514 Gehrke M.; Germany "GNEIS: A Portable Natural Language Explanation Component for Expert Systems" 520 Campioli A., L Lucchesi, F Turini; Italy "Spread views" Pen eva J., J Angelova; Bulgaria "End- User Interface to Improve Microcomputer DBMS Efficiency" 525 531 558 [16] M Musen, D M Combs, J.D Walton, E.H Shortliffe, L.M Fagan; OPAL: Toward the Computer-Aided Design of Oncology Advice System; in Proceedings of the 10th Annual Symposium on Computer Applications in Medical Care, IEEE, pp 43-52, 1986 [17] R.S Patil, O Senyk; Efficient Structuring of Composite Causal Hypotheses in Medical Diagnosis; in Proceedings of the 11th Annual Symposium on Computer Applications in Medical Care, IEEE, Washington D C., pp 23- 29; 1987 [18] Harry E Pople, Jr.; Heuristic Methods for Imposing Structure on lli-Structured Problems: The Structuring of Medical Diagnostics; in Szolovitz P., ed Artificial Intelligence in Medicine, Westview Press, pp 119-190; 1982 [19] ROCHE-Dictionary of Medicine; ed.: Hoffmann-La RocheAG; Urban & Schwarzenberg, Munich; 1987 (in German) [20] Singer, J.; Sacks, H.S.; Lucente, F.; Chalmers, T.C.: Physician Attitudes towards Applications of Computer Data Base Systems, in: JAMA, Vol 259, pp 1610-1614, 1983 [21] Stary, Ch.; Adolf, W.: Computer - Supported Therapy Decision Making for the Malignant Melanoma of the Skin (in German), in: GI Informatik-Fachberichte, Vol 127, eds.: Hommel, G.; Schindler, S., pp 482-496, Springer, Berlin, October 1986 [22] Gio Wiederhold, Robert L Blum, Michael Walker; An Integration of Knowledge and Data Representation; in On Knowledge Base Management Systems: Integration Artificial Intelligence and Database Technologies, Brodie, Mylopolous and Schmidt (Eds.), Springer; 1986 Symbolic computation in RL/l S van Denneheuvel K L Kwast P van Emde Boas Department of Mathematics and Computer Science University of Amsterdam F de Geus E Rotterdam Department of Medical Information Science University of Groningen Abstract This article discusses the use of RL/l to build quantitative models and the application of these models to assist in decision making A set of constraints constitutes a model Constraint models can be extended and modified easily and the representation of model knowledge is separated from its use The model knowledge is used by means of a constraint solver This allows the application of the same knowledge for different purposes In addition the knowledge representation allows the user to analyze relatioliGhips between variables and to infer new relationships that hold in special circumstances The rule language family RL Recently, systems have emerged that allow declarative constraint processing (see for example Mathematica [9], Bertrand [5] and eLP [3]) Declarative constraint processing is also pursued in the rule language family RL (see [6], [7] and [8]), where knowledge can be represented in three different types of rules: tabular rules, clauses and constmints Corresponding to these types of rules there are three areas of technology that support that style of knowledge processing in isolation: database systems, logic programming systems, and spreadsheets The main goal of RL is to integrate these three technologies in one knowledge base system In the current prototype RL/l we focus on the integration of a constraint solving subsystem with a relational database system; the associated language RL/l was presented previously at this conference [2] All examples presented in this paper were processed successfully by this RL/l prototype In RL/l a distinction is made between extensional objects and intensional objects Extensional objects correspond to base tables in the underlying relational database Intensional objects are objects whose relations can be materialized by evaluation of their definition An intensional object is defined by one or more rules, alternatively called clauses A rule typically consists of a rule head and a rule expression separated by the keyword WHEN The rule expression contains system defined primitives or ~nvocations of intensional and extensional objects, separated by the conjunctive AND operator Defining an intensional object with several rules expresses disjunction between the rules Queries in RL/l result in an answer relation which.consists of attributes and a (possibly empty) set of tuples Optionally the result of a query can be stored in an extensional table object (cf [2]): INFER (attribute-li~t rule-exp [ TO table ] WHEN The INFER query yields an answer relation with attributes equal to the attribute list between the INFER and WHEN keywords The processing of queries requires one or more invocations of the constraint solver The solver output is used to compile the query into a database request Symbolic queries work by presenting this intermediate solver output to the user directly They have the following format: SYMINFER attribute-list) WHEN rule-exp [ TO file ] With the 'TO file' option the computed symbolic answer can be stored in a text file Symbolic query commands produce the following answer: con= Condition set 501= Solution set = The solution set contains elements of the form a: with a: a wanted variable and t a term Wanted variables are variables that are to be eliminated from the constraint set The condition part states under what additional restrictions the obtained solution is valid The main goal of the constraint solver is to express the wanted variables in terms of known variables Known variables are determined To see if theory matches reality we apply 10 volt at the clamps of the circuit When the values of all resistors are 200 ohm, the circuit should be in balanced state and as a consequence no current is should flow through TS which indeed turns out to be the case: infer(vb,vc,il,i2,i5) vhen circuit(iO,vb,vc,O, il,i2,i3,i4,i5,200,200,200,200,200) vb vc il i2 i5 5.000 5.000 0.025 0.025 Figure 1: The bridge of Wheatstone module vheatstoneenumber:v:i:r:va:vb:vc: vd:ve:il:i2:i3:i4:i5:rl:r2:r3:r4:r5) clause ohmlavev,i,r) vhen v=i*r clause circuiteva,vb,vc,vd,il,i2,i3,i4,i5, rl,r2,r3,r4,r5) vhen ohmlaveva-vb,il,rl) and ohmlaveva-vc,i3,r3) and ohmlavevc-vd,i4,r4) and ohmlav(vb-vd,i2,r2) and ohmlav(vc-vb,i5,r5) and il +i5=i2 and i4+i5=i3 Figure 2: Representation in RL/l by the rule expression of the query and wanted variables by the attribute list of the query The capability to produce symbolic output is an important feature of the RL/l knowledge base system It allows the user to analyze relationships between variables and to infer new relationships that hold in special circumstances Symbolic computation Intensional objects are declarative representations of knowledge; therefore it is not known beforehand how the attributes of an object are invoked As an application in the domain of electronics consider the circuit given in Figlire and its RL/l representation in Figure In ancient history the bridge of Wheatstone was used to measure ohm resistance In the setup, Tl (or T2) is a variable resistor and T4 (or T3) the unknown resistor The component in the center is an ampere meter with resistance TS For the determination of the unknown resistor the variable resistor is adjusted so that no current flows through TS In this balanced state it can be shown that the Wheatstone relationship holds: If one of the resistors, say Tl, is taken unequal to the others, a small positive current flows through the center resistor: infer(vb,vc,il,i2,i5) vhen circuit(10,vb,vc,0, il,i2,i3,i4,i5,400,200,200,200,200) vb vc il i2 i5 3.846 4.615 0.015 0.019 0.003 Setting T2 instead of Tl to 400 ohm results in a nt'gative current through the center resistor, as expected: infer(vb,vc,il,i2,i5) vhen circuit(10,vb,vc,0, il,i2,i3,i4,i5,200,400,200,200,200) vb vc il i2 i5 6.153 5.384 0.019 0.015 -0.003 As mentioned before, the constraint solver tries to eliminate wanted variables from the constraint set and all wanted variables are returned in the solution set The solver also produces a condition set with expressions that only contain known variables A reduced model of the circuit can be inferred by giving particular values for all resistors of the circuit To obtain the reduced model the variables rl, r2, r3, r4 and TS are made wanted and all other variables known The values of the resistors are presented to the solver in equations The wanted variables are returned directly in the solution set and the condition set contains the reduced model for the circuit with the substituted resistor values (the '( *)' notation expands to all attributes of the involved object): syminfer(rl,r2,r3,r4,r5) vhen circuit(*) and known9(va,vb,vc,vd,il,i2,i3,i4,i5) and rl=200 and r2=200 and r3=200 and r4=200 and r5=200 con= -i3 + i4 + i5 = 0, il - i2 + i5 = 0, -200 * i5 - vb + vc = 0, 200 * i2 vb + vd 0, 200 * i4 - vc + vd 0, 200 * i3 - va + vc 0, 200 * il - va + vb 501= rl=200,r2=200,r3=200,r4=200,r5=200 Since the clamps Va and Vd are inputs of the circuit it should be possible, in the above reduced model where values are given for all resistors, to express all variables of 561 the circuit in terms of Va and Vd The solver is actively directed towards this goal by making only Va and Vd known: syminfer(vb,vc,i1,i2,i3,i4,i5) when circuit(va,vb,vc,vd,i1,i2,i3,i4,i5, 200,200,200,200,200) and known2(va,vd) 501= vb 0.50000 va + 0.50000 • vd, vc 0.50000 va + 0.50000 • vd, i1 0.00250 va - 0.00250 • vd, i2 0.00250 va - 0.00250 • vd, i3 0.00250 va - 0.00250 • vd, i4 0.00250 va - 0.00250 • vd, i5 Note that since the variables 7'1,7'2, 7'a, 7'4 and 7's not appear in the query, their solution is not present in the symbolic output as in the previous query Instead the values for the resistors were specified directly in the invocation of the object circuit If one of the resistors is taken unequal, the reduced model is no longer symmetric: syminfer(vb.vc.i1.i2.i3.i4,i5) when circuit(va,vb.vc.vd.il.i2,i3.i4.i5, 400.200.200.200.200) and knovn2(va.vd) 501= vb 0.38461 va + 0.61538 • yd vc 0.46153 • va + 0.53846 • yd 0.00153 • va - 0.00153 • yd i1 i2 0.00192 • va - 0.00192 • yd i3 0.00269 • va - 0.00269 • vd, 0.00230 • va - 0.00230 • yd i4 i5 0.00038 • va - 0.00038 • vd It might be quite interesting to see if the Wheatstone relationship for 7'4 can be inferred symbolically from the represented knowledge under the assumption that the circuit is in balanced state This assumption yields a reduced model We want to express 7'4 in terms of 7'1, 7'2 and 7'a so these latter variables are specified as knowns and 7'4 is specified as the only wanted variable As an extra constraint we know that since the circuit is in balanced state no current flows through 7's: syminfer(r4) when circuit(.) and i5=0 and known3(r1.r2.r3) 501= r4=r2.r3/r1 I To obtain this answer, the solver had to process the following constraint set: v a -vb=i1 *7'I, Va -V c =ia *7'3, v c -vd=i4 *7'4, = i2 * 7'2, Vc - Vb = is * 7's, il + is = i , i4 + is = i , is = Vb - Vd These equations are nonlinear by the occurrences of the terms il * 7'I, i2 * 7'2, i3 * 7'3, i4 * 7'4 and is 7's (note that the known variables 7'1, 7'2 and 7'3 are in fact not given as a value and therefore contribute to the nonlinearity of the constraint system) Quick observation of the constraint set shows a total count of 14 variables of which variables are known and the remaining 11 are unknown There are equations to express unknown variables in terms of known variables A simple calculation predicts that we might eliminate unknown variables by using the equations and as a consequence we are left with (= 11-8) unknown variables that can not be expressed in terms of known variables Our hope is that 7'4 is among the variables that can be expressed in terms of the known variables In fact as we know from elementary physics, 7'4 is the only variable that can be expressed in 7'1, 7'2 and r3 A medical application This section describes a physiological model of the human blood circulation and gas exchange It is a simple model and therefore a poor reflection of reality, but it serves well to demonstrate how a model represented in constraints can be used for decision support in medicine Variable names used in the model are abbreviated according to medical convention A summary of the properties of the variables used is shown in Figure The model consists of eleven equations; describe the systemic and pulmonary circulation and the remaining the gas exchange Each model equation is presented together with some explanatory text Figure shows a graph representation of the model VV = KV x In(CVP) A higher or lower venous volume (VV) results in a higher or lower central venous pressure (CVP) When VV gets larger, the veins are stretched and resist further filling The above formula is derived from the assumption that the change in CVP due t~ a change in VV, is proportional to CVP: 6CVP/6VV ~ CVP SV = KR x In(CVP) For the right side of the tJ-eart a similar formula can be assumed for the relationship between the stroke volume (SV) and the CVP This formula has been chosen to keep the model simple In reality the stroke volume increases initially, but it decreases again above a certain level of the central venous pressure This relationship is known as the Starling curve = SV KL x In(PLA) This relationship is the equivalent for the left side of the heart = PArtM - CVP CO x SVR Ohm's law for the systemic circulation The pressure difference between the mean a.rterial pressure As a comparison we ran this problem on a Mathematica system After 91 seconds the system responded with a formula consisting of 15 subcases, none of which contained the desired relationship "4 = T2 * Ta/Tl The RL/l system computed the result in seconds 562 ~ CONTINUOUS I'dEASUREABLE o INTERMITTENT I'dEASUREABLE Figure 3; Graph re prese ntation of the model (PArtM) and the CVP is equal to the cardiac output (CO) multiplied by the systemic vascular resistance (SVR) 11 RQ The respiratory quotient (RQ) is defined as VC0 divided by V0 When no changes take place in the total body content of carbon dioxide, it has a constant value of 0.8, since the metabolic rate of oxygen used is proportional to the quantity of carbon dioxide produced PAPM - PLA = CO x PVR The equivalent relationship for the pulmonary circulation CO = SV x HR The CO is the volume of one heart stroke multiplied by the heart rate (HR) Ca02 = 0.000134 x Hb x Sa02 The arterial oxygen concentration (Ca02) is proportional to the haemoglobin concentration (Hb) multiplied by the fractional arterial oxygen saturation (Sa02)' The constant 0.000134 is in dl/g CV02 = 0.000134 x Hb x SV02 The equivalent relationship for the venous system 9': V02 = (Ca02 - CV02) x CO The oxygen intake per minute (V0 2) is equal to the difference in oxygen concentration between the arterial and the mixed venous blood multiplied by the CO 10 VC0 = (CvC0 - CaC0 2) x CO The expired carbon dioxide production per minute (VC0 2) is supposed to be the difference in carbon dioxide concentration between the venous and arterial blood multiplied by the CO VC0 =~V0 Representation of the model The physiological model is represented in RL /1 by a single rule defining the intensional object s as shown in Figure The '*' notation in the head of a rule indicates that all variables occurring in the associated rule expression are declared as attributes This section discusses two ways in which a physiological constraint model can be used to assist anaesthetists in decision making; (1) the interpretation of measurements, and (2) the prediction of effects of treatment The anaesthetist tries to gain insight in the development of the physiological condition of a patient by monitoring the trend of physiological variables Unfortunately not all variables can be measured A physiological model, like the one in Figure 5, can be used to derive values for non measurable variables from the ones measured The number of physiological variables which can be measured has increased over the past few decades This development has raised a demand for computer systems that present all measured information in a coherent way (see e.g [lJ and [4]) The rapid development of new mea- 563 Abbreviation CO CVP CaC02 Ca02 CvC02 CvO, lIR lIb KL KR KV PAPM PArtM PLA PVR RQ Sa02 SvO, SV SVR VCO, VO, VV units a value ml/min mmHg ml/ml ml/ml ml/ml ml/ml B/min gr/dl mmHg/ml mmHg/ml mmllg/mi mmHg mmHg mmHg mmHg (ml/min) % % ml mmHg (nd/nl'in) ml/min ml/min ml Name Cardiac Output Central Venous Pressure Arterial CO2 Concentration Arterial O2 Concentration Venous CO Concentration Venous O Concentration Heart Rate Haemoglobin concentration Left Heart parameter Right Heart parameter Venous system parameter Mean Pulmonary Arterial Pressure Mean Arterial Pressure Left Atrium Pressure Pulmonary Vascular Resistance Respiratory Quotient Arterial O saturation Venous O saturation Stroke Volume Systemic Vascular Resistance Expiratory CO, Rate Inspiratory O Rate Venous Volume module phsystem(number:Ca02:CaC02:CvC02: Cv02:CO:CVP:HR:KL:KR:KV:PAPM:PArtM:PLA: PVR:RQ:Hb:Sa02:Sv02:SV:SVR:VC02:V02:VV) table continuous(PArtM,CVP,PLA,PAPM,HR, Sa02,VC02) table intermittent(CO,Hb,Sv02) table patient(KV,KR,KL) table unknown(VV,SV,SVR,PVR,Ca02,CaC02, Cv02,V02,CvC02,RQ) clause 5(*) when 1* 1*1 VV=KV*ln(CVP) 1* 2*1 and SV=KR*ln(CVP) 1* 3*1 and SV=KL*ln(PLA) 1* 4*1 and PArtM-CVP=CO*SVR 1* 5*1 and APM-PLA=CO*PVR 1* 6*1 and CO=SV*HR 1* 7*1 and Ca02=0.0001134*Hb*Sa02 1* 8*1 and Cv02=0.0001134*Hb*Sv02 1* 9*1 and V02=(Ca02-Cv02)*CO 1*10*1 and VC02=(CvC02-CaC02)*CO 1*11*1 and RQ=VC02/C02 close Figure 5: The physiological model in RL/l Figure 4: Properties of model variables ammHg = millimeter mercury, B = beats surement equipment also requires that a computer system is adaptable to new measurement possibilities Clearly this requirement is met by a declaratively represented model In our model physiological variables can be divided into four classes Continuously measurable variables: in the University Hospital of Groningen a computer system on the theatre stores average values for these variables at minute intervals [4J These sets of values can be used as input for interpretation Intermittently measurable variables: values of these variables can not be automatically measured, and the measurement methods available demand extra attention of the anaesthetist They are therefore performed only when necessary To determine bloodgasses for example, a sample of blood has to be taken The time to analyze a blood sample in the laboratory averages to 18 minutes Another example is cardiac output CO, which is measured by inject.ing cold fluid into the blood stream This introduces extra volume and ca.n therefore only be done a few times during an operation Patient parameters: variables whose values are characteristic for a particular patient and cha.nge so slowly that they can be assumed constant with respect to other variables during an operation Unknown variables: variables, that can be neither measured nor assumed to be constant In Figure tables are defined to group continuously measurable variables, intermittently measurable variables, patient parameters and unknown variables respectively 4.1 Interpretation Anaesthetists are interested in the trend of the cardiac output (CO) An incidental measurement of CO, combined with values for the continuously measurable variables HR, CVP and PLA, determines the patient parameters KR and KL: syminfer(KR,KL) when 5(*) and CO=5000 and PArtM=72 and CVP=5 and PLA=17 and PAPM=27 and HR=72 and Sa02=0.99 and VC02=200 501= KR = 43.148259 , KL = 24.510842 By symbolic computation KR and KL can be expressed directly in terms of continuously measurable variables and CO: syminfer(KR,KL) when 5(*) and continuous(*) and known1(CO) 501= KR CO * 1 In(CVP) * 1 HR, KL = CO * 1 In(PLA) * 1 HR 564 KR and KL are assumed to remain equal to the values computed for them since they are patient parameters The values for KL and KR are used to construct a reduced model from the original one (cf Section 2) Subsequent measurements for CVP, PLA and HR can be used to compute the CO As shown below an estimate for the actual CO results syminfer(CO) when s(*) and PArtM=72 and CVP=5 and PLA=17 and PAPM=27 and HR=72 and Sa02=0.99 and VC02=200 and KR=43 and KL=24.5 and KV=2750 con= FALSE sol= CO = 4997.788344 The condition contains FALSE, which indicates that an inconsistency is detected The cause of this inconsistency can be ascertained by making both the continuously measurable variables and patient parameters known instead of supplying their values as in the previous query: syminfer(CO) when s(*) and continuous(*) and patient(*) con= KR * In(CVP) KL * In(PLA) sol= CO = HR * KL * In(PLA) The condition between KL, KR, CVP and PLA is not satisfied by the measured values 4.2 Prediction As an example of prediction of the effect of a treatment, consider a drug that increases the rigidity of the heart muscle This results in a decrease of KL and KR So our question is what the consequences are when KL and KR decrease by say 10% To describe the qualitative consequences, an increase in variable x is denoted as xl and a decrease as xl The qualitative consequences of decreases in KR and KL are known to be SV1, COL, HRl, PARTM1, PAPMl, SVRl, PVR!, CVPL PLAl, CvC0 21, Cv021, Sv021 As CVP and PLA decrease, the relative decrease in SV will be bigger than in KL and KR The quantitative model given is not refined enough to quantify all the changes listed above Therefore some assumptions have to be made When it is assumed that the continuously measurable variables keep their actual value, the effects of a 10% decrease in KL and KR are predicted by the following RL/1 query: syminfer(CO,SV) when s(*) and PArtM=72 and PLA=17 and PAPM=27 and HR=72 and Sa02=0.99 and VC02=200 and KR=43*0.9 and KL=24.5*0.9 and KV=2750 sol= CO 4498.009505, SV = 62.472354 Conclusion The language RL/1 provides a framework in which algebraic constraints can be combined with relational databases The prototype implementation RL/1 contains a constraint satisfaction module that generates symbolic solutions, which were used to infer reduced models and analyze relationships between variables We used symbolic computation in a medical application to interpret patient measurements, to predict treatment effects and to trace inconsistencies between measurements Substitution of measured variable values results in a reduced model When it is assumed that variables of a particular class have kept the values that were last found for them, values of other variables can be inferred using the reduced model References [lJ Annejet P Meyler, Automation in Anesthesia, a Relief?; evaluation of a data acquisition and display system, thesis of the university of Eindhoven, The Netherlands, ISBN 90-9001441-1, (1986) [2J van Denneheuvel, S & van Emde Boas, P., The rule language RL/l, A M Tjoa & R Wagner (eds.), Database and Expert Systems Applications, Proc of the International Conference, Vienna, Austria, Aug 1990, Springer Verlag Wien, 381-387, (1990) [3J Jaffar, J & Michaylov, S., Methodology and Implementation of a CLP System, Logic Programming, Proc of the Fourth Int Conf., Ed Jean-Louis Lassez, 196-218, (1987) [4J Karliczek, G F., Geus, F A., Wiersma, G., Oosterhaven, S & Jenkins, 1., CarDia, a ~omputer systemfol' automatic documentation in anesthesia, Int Clin Monitoring and Computing 4, 211-221, (1987) [5J LeIer, Wm., Constraint Programming Languages: Their Specification and Generation, Addison-Wesly Series in Computer Science, Addison- Wesly, (1988) [6J Van Emde Boas, P., RL, a Language for Enhanced Rule Bases Database Processing, Working Document, Rep IBM Research, RJ 4869 (51299), (1986) [7J Van Emde Boas, P., A semantical model for the integration and modularization of rules, Proceedings MFCS 12, Bratislava, Springer Lecture Notes in Computer Science 233, 78-92, (1986) [8J Van Emde Boas, H & van Emde Boas, P., Storing and Evaluating Horn-Clause Rules in a Relational Database, IBM J Res Develop 30 (1), 80-92, (1986) [9J Wolfram, S., Mathematica, A System for Doing Mathematics by Computer, Addison-Wesfy, (1988) INEKS - an Information System for Medical Application Roland Zimmerling University of Oldenburg Computer Science Departtnent P.O Box 2503 D-2900 Oldenburg Abstract This Paper describes the development of the information system INEKS *) - INEKS for the Stadtische Kliniken Oldenburg After building Informationssystem fUr l