Case-Based Reasoning in Decision Support Systems

Case-Based Reasoning (CBR) is a memory-based approach that supports decision making by retrieving and adapting solutions from a collection of past cases. Instead of relying on abstract models, CBR systems draw on concrete examples of previous problem–solution pairs, enabling rapid responses in domains where expert knowledge is critical. The typical CBR cycle comprises retrieval of similar cases, reuse of their solutions, revision to suit the new context and retention of the new experience. This methodology has found global application in fields as diverse as medical diagnosis, emergency response, manufacturing and environmental management. Recent advances focus on enhancing case representation through knowledge graphs, refining similarity measures via machine-learning techniques and developing hybrid frameworks that integrate CBR with reinforcement learning or rule-based engines. Such innovations address longstanding challenges of efficient case retrieval in large repositories, dynamic adaptation of solutions to novel circumstances and ongoing maintenance of case libraries to ensure relevance and manage computational cost. By learning from experience and continuously evolving, modern CBR systems offer adaptable, data-driven decision support that balances academic rigour with practical utility.

Research from Nature Portfolio

No recent Nature Portfolio content available.

Research from all publishers

In the realm of search and rescue operations, a recent applied artificial intelligence study introduced a modular decision support framework that combines term-weighting retrieval methods (BM25, TF-IDF) with clustering algorithms for case matching. Hierarchical agglomerative clustering utilising BM25 achieved superior coherence and was validated by domain experts for resource allocation under uncertain conditions. In planning scenarios, another investigation formalised the problem of maintaining a plan library within CBR planners, proposing both offline reduction techniques and online growth-limiting policies. These strategies balance the trade-off between library size and retrieval efficacy, yielding faster planning times without degrading solution quality. In industrial manufacturing, a hybrid intelligent system was developed for integrated bogie welding decisions. This system blends CBR, rule-based reasoning and a knowledge graph of domain causality to process structured documents and extract decision rules, achieving a corrected accuracy of 0.947 in engineering trials and demonstrating clear value for complex production workflows.

Case-Based Reasoning in Decision Support Systems publication trend

The graph below shows the total number of articles in case-based reasoning in decision support systems across all publications each year (not limited to Nature Index journals).

Technical terms

Case library: A structured repository of past problem–solution records used for case retrieval.

Similarity measure: A quantitative metric that evaluates the closeness between a new problem and stored cases.

Case adaptation: The process of modifying a retrieved solution to fit the specifics of a new case.

Knowledge graph: A semantic network that organises entities and relationships to enrich case representation and retrieval.

Maintenance policy: A strategy for updating and pruning the case library to maintain relevance and computational efficiency.

References

1. Knowledge Graphs to Accumulate and Convey Knowledge from Past Experiences in Search and Rescue Planning and Resource Allocation. Applied Artificial Intelligence (2024).
2. Maintenance of Plan Libraries for Case-Based Planning: Offline and Online Policies. Journal of Artificial Intelligence Research (2023).
3. Hybrid Decision-Making-Method-Based Intelligent System for Integrated Bogie Welding Manufacturing. Applied System Innovation (2023).
4. Learning similarity measures from data. Progress in Artificial Intelligence (2019).
5. Scenario-Based Marine Oil Spill Emergency Response Using Hybrid Deep Reinforcement Learning and Case-Based Reasoning. Applied Sciences (2020).