Efficient Forward Chaining for Declarative Rules in a Multi-Agent Modelling Language

Contents

Contents

Abstract

1 - Introduction

2 - Motivations for strictly declarative rules

3 - Representation of knowledge in SDML

3.1 - Clauses

3.2 - Databases

3.3 - Rulebases

3.4 - Models

4 - Firing rules

4.1 - Forward chaining

4.2 - Compilation

4.3 - Dependencies and partitioning

4.4 - Optimising cycles in rulebases

4.5 - Backward chaining

5 - Assumption handling

5.1 - Creating assumptions

5.2 - Assumption tags

5.3 - Resolving assumptions

5.4 - Optimisations

5.4.1 - Partitioning

5.4.2 - Extending dependency graphs

5.4.3 - Ordering within partitions

5.4.4 - Invalid combinations of assumptions

5.4.5 - Validities

5.5 - Random and arbitrary decisions

6 - Performance

6.1 - Benchmarks

6.2 - Measurements

7 - Comparisons with other approaches

7.1 - Logic

7.2 - Truth maintenance systems

7.3 - Other declarative rule-based systems

8 - Conclusions

Acknowledgements

References

Generated with CERN WebMaker