Reasoning methods for ME-maps — A CSP based approach

Abstract

Know-how refers to the knowledge of how to achieve objectives effectively and efficiently. Mapping this knowledge helps in understanding domains, learning about problems and solutions, identifying potential gaps and places for improvements, and reasoning about the existing knowledge. Recently, we developed and formalized an approach called ME-MAP for mapping out know-how. Checking the correctness properties of such maps is essential to ensure their qualities and grantee their practical usage. Therefore, it is critical to enable automation of reasoning methods. This paper presents three reasoning methods for ME-maps: (1) checking consistency; (2) creating legal instances; and (3) extracting a “core map” for a given ME-map. The methods are based on translating ME-maps into a Constraint Satisfaction Problem (CSP) and then using an off-the-shelf CSP solver for verifying the desired properties.