A Casebased Strategy for Solution Synthesis among Cooperative Expert Systems.pdf

A Casebased Strategy for Solution Synthesis among Co op erative Exp ert Systems Minjie Zhang Scho ol of Computer Information and Mathematical Sciences Edith Cowan University WA Australia mzhangcowaneduau Abstract Most synthesis strategies for solution synthesis in distributed exp ert systems DESs are develop ed based on analytic metho ds or in ductive metho ds A synthesis strategy presented in this pap er is devel op ed by casebased reasoning based on an analogical metho d In this pap er the basic frame work of the strategy is intro duced and prelimi nary results are demonstrated Furthermore this strategy is compared with related work Intro duction Synthesis of solutions has b een b ecome one of critical issues among co op erative exp ert systems for more than ten years Some synthesis strategies have b een develop ed in distributed exp ert systems DESs such as These strategies were mainly based on the mathematical analysis of the characteristics of the inputs multiple solutions from ESs In recent years we have develop ed some synthesis strategies using neural networks These strategies were based on a numb er of samples with b oth inputs and corresp onding outputs nal solutions after synthesizing corresp onding input Such samples are called p atterns In current literature there are two metho dologies used for designing synthesis strategies to solve synthesis problems They are analytic metho ds and inductive metho ds An analytic metho d is a metho dology which can b e used to dene a synthesis strategy from inputs to outputs by analyzing the characteristics of input These characteristics may include relationships among ori