高性能计算导论课件智能规划简介for2015master -1.ppt

Thank you! The main differences between Planning Based on the Model-Checking and classic planning： The states are no longer represented as atomic propositions， but represented as propositional sets. So finite state machine has contained domain knowledge. BDD (Binary Decision Diagrams): Planning Based on the Model-Checking adopts the logic representation structure of BDD. BDD is a type of data structure using to represent boolean functions. It provides a valid, normal representation method for boolean functions，and is widely used in Model-Checking domain，especially in protocol verification. BDD is a type of data structure based on boolean functions，and has more often applied to deterministic planning and non-deterministic planning, such as BDDPLAN, PROPPLAN and MIPS. The key factor about adopting BDD： When using BDD to represent two predicate formulas，two BDD normal forms are logically equal if and only if they are the same BDD normal form. See WORD file for examples of planning as model checking. PART Ⅵ Applications of Planning Industry: Planning is a research domain with very strong utility in AI research. For example, Job shop scheduling problems consider : under the condition of finite process resources (lathe, planer, drill press), according to the known required job order of work pieces, arrange the manufacturing of the whole workshop so as to make as less as possible the needed time of finishing processing all work pieces and the waiting time of every machine tool . Industry: Another typical job shop scheduling problem considers carrying freight between different locations under the precondition of finite power trucks. The output of planning is a vehicle running schedule, making that automobiles transport with full load as much as possible, and that the conditions of empty vehicle transportation and vehicle left unused are as few as possible. This will also certainly bring considerable benefit to transportation corporations. Aviation and Aerospace :