搜 索 算 法（Search algorithm）.docVIP

搜 索 算 法（Search algorithm） Search algorithm The search algorithm is the purpose of a problem exhaustive use of high-performance computer to some or all of the possible, to find a solution A method. The search process is actually based on the initial conditions and the expansion of the rules of the tree structure and find an answer process in line with the objectives of the state of the node. All of the search algorithm from the final algorithm implementation, can be divided into two parts: control structure and system, and all the Optimization and improvement method is mainly accomplished by modifying its control structure. Now the main control of the structure are discussed, so the Department of The agreement is as follows: Function ExpendNode (Situation:Tsituation; ExpendWayNo:Integer): TSituation; Said Sitution on the node state given by the ExpendWayNo extension expansion rules, and returns the extended state. (the algorithm description language for Pascal.) The first part is the basic search algorithm A backtracking algorithm Backtracking algorithm is one of the most basic algorithm in all search algorithm, which uses a dead turned thought as its control structure, To construct the answer tree method is equivalent to the first traversal, can be used to find all solutions and the optimal solutions or solutions. The specific algorithm is described as follows: Non recursive algorithm [] Type Node (node type) = Record Situtation:TSituation (the current node); Way-NO:Integer (the number of expansion rules have been used.); End Var List (back table): Array[1..Max (maximum depth)] of Node; POS (the current expansion node number: Integer); Init List-0; Pos-1; The initial state of List[1].Situation-; Main Program While (pos0 (a road) (and) [do] did not reach the target) Begin If pos=Max then (data overflow, jump out of the main program); List[pos].Way-NO:=List[pos].Way-No+1; If (List[pos].Way-NO=TotalExpendMethod) then (if you have not used the extension rule) Begin If (