Prototype Meta-Language and Automated Translator for Decision Analysis ProblemFormulation

摘要

This thesis research effort develops a prototype meta-language for decisionanalysis problem description and automated software to translate the meta-language into a format useable by decision analysis software solvers. MELDA, the metalanguage developed, makes use of special symbols to provide a succinct language for the automation environment yet provides the means for user comprehension. Complete syntax and implementation rules are included. The MELADA translator software developed, DAT, stores the data given by MELADA in a file using prototype standard format, DASF, designed to hold all required data to solve a problem. DAT provides error checking and a message output file. TREESOLVER, an add on program to solve decision trees stored in DASF format was designed. The program uses the external DASF file directly to bypass size limitations of internal memory. All software is written in Turbo Pascal 5.5, is compatible with any MS-DOS computer system and is free to government users in source code format, including a user's manual. Finally, several example problems are solved using MELADA, DAT, DASF, and TREESOLVER to validate their viability.