Optimization And Implementation Of Epistemic Specification Inference Algorithm

Epistemic Specification improves expressive and reasoning ability via epistemic operators. Traditional epistemic specification has inefficient solver and cannot express subjective probability. This paper extends traditional epistemic specification, designed an epistemic specification with probability. Based on this logic programming language, we focus on the design, optimization and implementation of the reasoning algorithm. Besides, we present a solving method for conformant planning problem which is based on epistemic specification. We adopted several conformant planning problem to verify it.The main researches of this thesis include:1) Design an epistemic specification with probability, define its syntax and semantics, and consider its property under the semantics of world views.2) Design the framework of a solver for Epistemic Specification with probability. Design several optimize algorithms during grounding and reasoning period to improve solver’s efficiency.3) Design a general algorithm to solve probabilistic conformant planning problem.4) Present some benchmark problems to verify epistemic specification reasoning algorithm.The main contributions of this thesis include:1) Designing a new logic programming language which adds probability to epistemic operator.2) Designing a complete solver framework which includes front-end and back-end.3) Designing an epistemic specification reasoning algorithm based on answer set solver.4)Based on the language of ASPPK, we give a modeling method to solve probabilistic conformant planning problems with threshold.5) Presenting several probabilistic conformant planning problems (toilet and bomb, slippery gripper, blocks world) to verify correctness and efficiency of the reasoning algorithm.