A Survey of Multi-agent Systems and Case-Based Reasoning Integration

Abstract

A Case-Based Reasoning (CBR) is defined as an Artificial Intelligent technique that has a problem-solving ability by reasoning previously experienced cases to solve new cases. The CBR is considered as one of the successful techniques that are applied in a widespread of problem-solving tasks and domains. However, the CBR in unknown or poorly archived cases suffers from uncertainty and imprecision. Additionally, the CBR is inefficient in performing partial reasoning and revision to distributed and dynamic problems. These problems entail flexible problem-solving and management architecture. The researchers integrate Multi-agent-systems (MAS) within the CBR to increase the ability of the CBR to solve problems that require agents' abilities such as interaction, autonomy, and flexibility. Consequently, in this paper, we have surveyed various techniques and methods that integrate MAS in CBR (CBR-MAS) to solve different challenges in different domains. The paper outcomes two main approaches of CBR-MAS: a number of agents are integrated on the reasoning steps of the CBR cycle or a CBR or CBRs is integrated on the run cycle of an individual agent or a MAS.