Expert Systems

Unlike traditional procedural systems that integrate algorithms and knowledge within a single program, expert systems distinguish between domain-specific expertise and the procedural methods used to execute the program. An expert system consists of three primary components:

1. Knowledge Base: This module stores domain expertise as a collection of simple if-then rules, serving as a repository of knowledge accumulated from various experts.
2. Working Memory: This component holds data inputs and monitors the progress of deductions made by the inference engine.
3. Inference Engine: This is a computer program that operates in a loop. During each cycle, it evaluates the facts in the working memory against the rules. When a rule’s conditions are met, the rule can generate new facts, modify existing ones, or halt the loop. The loop stops automatically if no more rules can be triggered with the current facts.

This separation of components offers several advantages. It enables individuals without programming expertise, such as doctors, traders, underwriters, or compliance experts, to contribute their knowledge to the knowledge base. Additionally, this setup simplifies the enrichment and maintenance of the expert system, as knowledge can be easily added, updated, or removed from the knowledge base without altering the underlying code.

Business Rules Management System

A Business Rule Management System (BRMS) allows organizations to centralize, manage, and automate complex decision-making processes. Technically, a BRMS includes several key components:

1. Rule Repository: This serves as the storage for business rules, where they are defined, organized, and efficiently stored.
2. Rule Authoring Environment: This provides user-friendly interfaces for creating, modifying, and testing rules, supporting rule development by business users or analysts.
3. Rule Coherence Module: This crucial component prevents contradictions between rules. It assesses new or modified rules to ensure they do not conflict with existing ones, avoiding confusion or unintended consequences in decision-making.
4. Inference Engine: This computational module executes the defined rules, processing and applying them to make decisions or automate actions within business processes.
5. Rule Governance: This system manages rule versions, tracks changes, and ensures compliance with regulatory requirements and internal standards. It maintains a history of rule modifications and their impacts.
6. Analytic Tool: This tool monitors and analyzes the effects of rules on business processes, providing insights and data-driven information. It includes “What-If” and “What-If-Not” simulation capabilities, helping organizations make informed decisions and optimize their rules to achieve desired outcomes while mitigating unintended consequences.

One of the main advantages of this system is its flexibility, allowing non-technical users to easily add, update, or remove rules. It’s like having a rulebook that adapts to your business needs, making it easier to maintain and modify rules as circumstances change.

However, the inference logic for business rules is similar to that of early expert systems and has limitations. A significant portion of human expertise cannot be fully encapsulated within concise if-then-else rules. Even when it seems feasible, it is challenging to extract and transfer the comprehensive knowledge of a proficient expert to a computer.