Java 霍普克罗夫特';s算法&DFA最小化
我想实现Hopcroft算法来最小化DFA。到目前为止,我可以删除无法访问的状态。问题是我不理解这个算法。我不知道如何实施它。有人能解释一下吗?或者扩展算法,使其更易于实现。我根本不了解算法的以下部分:Java 霍普克罗夫特';s算法&DFA最小化,java,algorithm,computation-theory,Java,Algorithm,Computation Theory,我想实现Hopcroft算法来最小化DFA。到目前为止,我可以删除无法访问的状态。问题是我不理解这个算法。我不知道如何实施它。有人能解释一下吗?或者扩展算法,使其更易于实现。我根本不了解算法的以下部分: let X be the set of states for which a transition on c leads to a state in A for each set Y in P for which X ∩ Y is nonempty and Y \ X is non
let X be the set of states for which a transition on c leads to a state in A
for each set Y in P for which X ∩ Y is nonempty and Y \ X is nonempty do
replace Y in P by the two sets X ∩ Y and Y \ X
算法如下:
到目前为止,我已经实现了什么(写得很糟糕,完成后会进行清理):
package-dregout;
导入java.io.*;
导入java.util.ArrayList;
导入java.util.Collections;
导入java.util.Iterator;
公共级dfamin{
//保存文件中数据的全局变量
纽姆州、纽马尔哈贝茨州、纽姆芬纳州的私营企业;
专用字符字母[];
私人国际财务报表[];
私有int[][]可传递;
/**
*@param args
*@抛出异常
*@throws NumberFormatException
*/
公共静态void main(字符串[]args)引发NumberFormatException,IOException{
国际新州、新墨西哥州、新墨西哥州;
字符字母表[];
国际最终状态[];
int[][]可传递;
/*
*输入文件格式:NumberOfState numberofAlphabets转换1转换2…NumberOfFinalState最终状态
*例如:
* 8 2 1 5 6 2 0 2 2 6 7 5 2 6 6 4 6 2 2 2 6
* 5 2 0 1 0 1 3 4 3 4 2 4 3 0 2 3
* 8 2 1 0 0 2 3 1 3 0 3 5 6 4 5 6 6 3 1 3
* 9 2 1 4 2 5 3 7 4 7 5 8 6 1 7 1 8 2 0 4 3 2 5 8
*/
//获取文件名并打开一个流来读取它
FileInputStream fileStream=newfileinputstream(“/path/to/file”);
BufferedReader br=新的BufferedReader(新的InputStreamReader(fileStream));
//存储文件中的每一行
弦线;
//从文件中读取每一行
而((line=br.readLine())!=null){
//从每行读取单间隔数据
String[]splittedLine=line.split(“”);
//从这行读numStates,nummalphabets
numStates=Integer.parseInt(splittedLine[0]);
numAlphabets=Integer.parseInt(splittedLine[1]);
//对于(int a=0;a如果无法单独从接受/拒绝行为中分辨出DFA所处的状态,则调用两个DFA状态等效。对于每种语言,接受该语言的最小DFA没有等效状态
Hopcroft的DFA最小化算法通过计算未最小化DFA状态的等价类来工作。这种计算的核心是一个迭代,在每一步中,我们都有一个比等价更粗糙的状态分区(即,等价状态总是属于同一个分区集)
最初的划分是接受状态和拒绝状态。很明显,它们是不等价的
假设我们在当前分区的同一集合中有状态q1和q2。如果存在符号σ,则转移函数为δ,使得δ(q1,σ)和δ(q2,σ)在不同的划分集合中,通过粗糙度不变量,我们知道这些状态是不等价的,也就是说,存在一个字符串x,使得delta(delta(q1,sigma),x)和delta(delta(q2,sigma),x)在是否接受/拒绝方面不同。但是delta(delta(q1,sigma),x)=delta(q1,sigma x),因此字符串sigma x区分q1和q2。您引用的逻辑是适当地分割一个分区集
正确性证明中有趣的部分是,当第2步不可能时,我们得到了真正的等价类
伪代码看起来比这更复杂,因为它关系到我们找到步骤2的应用程序的效率。阅读原始文章中给出的示例可能有助于将此算法可视化
在中指定该算法之前,我们通过一个示例对该算法进行了说明
详细信息。广泛使用列表处理来减少
计算时间
如果不画出DFA并在白板上浏览它,很难以直观的方式解释算法。就实现而言,它是最好的(以我的经验)将伪代码作为注释放入,并编写代码来实现每一部分。对于这个特定的算法,使用实际的数学风格的集合类(支持并集和交集等)非常有帮助,这样您就可以逐字实现(请注意,这不一定是最有效的方法,但它可以让您更轻松地翻译伪代码)。谢谢。我正在使用此逻辑实现我的代码,希望它是正确的!
package dRegAut;
import java.io.*;
import java.util.ArrayList;
import java.util.Collections;
import java.util.Iterator;
public class dfamin {
// Global variables to hold data from the file
private int numStates,numAlphabets,numFinalStates;
private char alphabets[];
private int finalStates[];
private int [][] transitionTable;
/**
* @param args
* @throws IOException
* @throws NumberFormatException
*/
public static void main(String[] args) throws NumberFormatException, IOException {
int numStates,numAlphabets,numFinalStates;
char alphabets[];
int finalStates[];
int [][] transitionTable;
/*
* INPUT FILE FORMAT: numberOfStates numberofAlphabets transitions1 transtions2 ... numberOfFianlStates FinalState(s)
* Example:
* 8 2 1 5 6 2 0 2 2 6 7 5 2 6 6 4 6 2 2 2 6
* 5 2 0 1 0 1 3 4 3 4 2 4 3 0 2 3
* 8 2 1 0 0 2 3 1 3 0 3 5 6 4 5 6 6 3 1 3
* 9 2 1 4 2 5 3 7 4 7 5 8 6 1 7 1 8 2 0 4 3 2 5 8
*/
// Take file name and open a stream to read it
FileInputStream fileStream = new FileInputStream("/path/to/file");
BufferedReader br = new BufferedReader(new InputStreamReader(fileStream));
// Store each line from the file
String line;
// Read each line from file
while((line = br.readLine()) != null){
// Read single spaced data from each line
String [] splittedLine = line.split(" ");
// Read numStates,numAlphabets from the line
numStates = Integer.parseInt(splittedLine[0]);
numAlphabets = Integer.parseInt(splittedLine[1]);
//for(int a=0;a<numAlphabets;a++){
//alphabets[a] = '0';
//}
transitionTable = new int[numStates][numAlphabets];
int tt= 2;
// Loop thorough the line and read transition table
for(int row=0;row<numStates;row++){
for(int col=0;col<numAlphabets;col++){
transitionTable[row][col] = Integer.parseInt(splittedLine[tt]);
tt++;
}// End of for-loop to go thorough alphabets
}// End of for-loop to go thorough states
// Read number of final states
numFinalStates = Integer.parseInt(splittedLine[2+numStates*numAlphabets]);
//System.out.println(numFinalStates);
// Read final states
int z=0;
finalStates = new int[numFinalStates];
int start = 3+numStates*numAlphabets ;
int end = (3+(numStates*numAlphabets))+numFinalStates;
for(int fs=start;fs<end;fs++){
finalStates[z] = Integer.parseInt(splittedLine[fs]);
//System.out.println(finalStates[z]);
z++;
}// End of for-loop to read all final states
dfamin x = new dfamin(numStates,numAlphabets,numFinalStates,finalStates,transitionTable);
x.minimizer();
System.out.println(x);
}// End of while-loop to read file
// Close the stream
br.close();
}
dfamin(int nS,int nA,int nFS,int fS[], int [][] tT){
numStates = nS;
numAlphabets = nA;
numFinalStates = nFS;
//alphabets = a;
finalStates = fS;
transitionTable = tT;
}// End of DFAMinimizer constructor
/*
* A method to minmize the dfa
*/
public void minimizer(){
// Remove unreachable States
ArrayList<Integer> reachableStates = reachableStates(numStates, numAlphabets,transitionTable);
// Store all final states
ArrayList<Integer> fStates = new ArrayList<Integer>();
// Loop thorugh finalStates array and transfer its data to array list
for(int fs:finalStates){
fStates.add(fs);
}// End of for-loop
// Store all non final states
ArrayList<Integer> nonFStates = new ArrayList<Integer>();
// Store only non final states in nonFstates
nonFStates = nonFinalStates(reachableStates,fStates);
//TODO: IMPLEMENT HOPCROFT's ALGORITHM
}// End of minimizer method
/*
* unreachableStates - A method to find unreachable states of a DFA
*
*/
public ArrayList<Integer> reachableStates(int numStates, int numAlphabets, int [][] transitionTable){
// Initialize a list to hold temporary list of states in it
ArrayList<Integer> reachableStates =new ArrayList();
ArrayList<Integer> newStates = new ArrayList();
// Start from the state zero
reachableStates.add(0);
newStates.add(0);
// Temporary array to hold reachable states
ArrayList<Integer> temp = new ArrayList();
// Loop until there is data in newStates
do{
// Empty temp array
temp.clear();
// Loop thorough all the states, and check for {p such that p=δ(q,c)};
for(int j=0;j<newStates.size();j++){
for(int i=0; i<numAlphabets;i++){
// If found add it to the temp set
temp.add(transitionTable[newStates.get(j)][i]);
} // End of for-loop to go thorough all characters
}// End of for-loop to go thorough all elements of the newStates array list
// Clear newStates list
newStates.clear();
// Add the elements that are in temp, but are not in reachableStates to newStates
// new_states := temp \ reachable_states;
for(int z=0;z<temp.size();z++){
for(int z1=0; z1<reachableStates.size();z1++){
// If the state was already present, don't add
if(temp.get(z) == reachableStates.get(z1)){
break;
}
if(temp.get(z) != reachableStates.get(z1) && z1 == reachableStates.size()-1){
// Only from temp to newstates if its not in reachablestates currently
newStates.add(temp.get(z));
}
}// End of for-loop to go thorough all reachableStates elements and check if a match
}// End of for-loop thorugh all temp states
// If newStates list is not empty then add it to the reachableStates
if(!newStates.isEmpty()){
// Add the newStates elements to reachable states
for(int y=0;y<newStates.size();y++){
//System.out.printf("newStates:%d newStatesSize:%d in %d",newStates.get(y),newStates.size(),y);
reachableStates.add(newStates.get(y));
}
}
}while(!newStates.isEmpty());
reachableStates = removeDuplicate(reachableStates);
return reachableStates;
}// End of unreachableStates method
/*
* removeDuplicate - a function to remove duplicate entries from an ArrayList
*
*/
ArrayList<Integer> removeDuplicate(ArrayList<Integer> input){
// Remove duplicate entries from reachableStates list
// Compare the first index, with all other indexes, compare the second with all other indexes
for(int i=0;i<input.size()-1;i++){
for(int j=i+1;j<input.size();j++){
// If dupblicate states remove it
if(input.get(i) == input.get(j)){
input.remove(j);
}
}
}// End of for-loop to remove duplicate entries from reachableList
// Sort the list before returning
Collections.sort(input);
// Return the list
return input;
}// End of removeDuplicate method
/*
* nonFinalStates - a method to return an array list of nonfinal states, given all and final states
*
*/
ArrayList<Integer> nonFinalStates(ArrayList<Integer> allStates, ArrayList<Integer> finalStates){
// All non final States
ArrayList<Integer> nonFinalStates = new ArrayList<Integer>();
// Loop thorough allStates, and compare each state with the list of finalstates
for(int i=0; i<allStates.size();i++){
// Loop thorough list of final states
for(int j=0; j<finalStates.size();j++){
// If a state is final state
if(allStates.get(i) == finalStates.get(j)){
// Then remove it from the list
allStates.remove(i);
}
}// End of for-loop to travers finalstates
}// End of for-loop to traverse allstates
return nonFinalStates;
}
// returns a string that is compatible with our input file specification
public String toString() {
StringBuffer buf = new StringBuffer();
//buf.append(" "+ numStates +" ");
//buf.append ( numAlphabets + " " );
buf.append("Transition Table: ");
for ( int i = 0; i < numStates; i++ ) {
for ( int j = 0; j < numAlphabets; j++ ) {
buf.append ( " "+ transitionTable[i][j] + " " );
}
}
buf.append ( "Number of Final State(s): "+numFinalStates + " Final State(s): " );
for ( int i = 0; i < numFinalStates; i++ )
buf.append ( finalStates[i] + " " );
return buf.toString();
}
}