操作系统进程调度算法(Java 实现)

FCFS(First Come First Server，先来先服务)

这是最简单，最基本的算法，它的思想非常简单，就是按照进程到来的时间顺序，逐个分配 CPU 资源
优点：简单，方便
缺点：效率低，资源利用率低

/**
  * CPU 占用情况
  * 1: 空闲
  * 0: 正被占用
  */
 static int CPU = 1;
 /**
  * 等待队列长度
  */
 static final int MAXLEN = 10;

 /**
  * 先来先服务算法
  * @param processes
  */
 public static void FCFS(List processes){
     int count = processes.size();
     int time = 0;
     int[] waitQueue = new int[MAXLEN];
     int front = 0;int tail = 0;
     int running = 0;
     System.out.println("-------------FCFS算法-------------");
     if (count <= 0){
         System.out.println("无可用进程");
         return;
     }

     while (count > 0){
         System.out.print("第 " + time + " 秒: ");
         for (int i = 0; i < processes.size(); i++){
             if (processes.get(i).isAlive() && processes.get(i).getInTime() == time){
                 System.out.print("进程 " + i + " 到来 ");
                 waitQueue[tail] = i;
                 tail = (tail+1) % MAXLEN;
             }
             if (processes.get(running).isAlive() && processes.get(running).getCount() == 0){
                 System.out.print("进程 " + running + " 结束运行 ");
                 processes.get(running).setAlive(false);
                 processes.get(running).setEndTime(time);
                 count--;
                 CPU = 1;
             }
         }
         if (CPU == 1 && front != tail){
             running = waitQueue[front];
             front = (front+1) % MAXLEN;
             System.out.print("进程 " + running + " 开始运行");
             int temp = processes.get(running).getCount();
             temp--;
             processes.get(running).setCount(temp);
             CPU = 0;
         } else if (CPU == 0){
             int temp = processes.get(running).getCount();
             temp--;
             processes.get(running).setCount(temp);
         }
         time++;
         System.out.println();
     }
     System.out.println("---------------------------------");
     ShowResult(processes);
 }

SJF(Short Job First，短作业优先)

按照进程预计需要的运行时间，按照从小到大分配资源
优点：简单进程执行速度快
缺点：无法准确预估运行时间，容易造成长进程饥饿

/**
 * 短作业优先算法
 * 就是在 FCFS 算法中加入对 waitQueue 等待队列按照运行时间排序
 */

//改变部分
...

for (int i = 0; i < processes.size(); i++){
    if (processes.get(i).isAlive() && processes.get(i).getInTime() == time){
        System.out.print("进程 " + i + " 到来 ");
        waitQueue[tail] = i;
        tail = (tail+1) % MAXLEN;
        length++;
        /**
         * 对等待队列按进程运行时长按从小到大排序
         */
        for (int x=front, z=0; z < length; x=(x+1)%MAXLEN, z++){
            for (int y=x+1, q=0; q < length-x; y=(y+1)%MAXLEN, q++){
                if (processes.get(waitQueue[x]).getCount() > processes.get(waitQueue[y]).getCount()){
                    int t = waitQueue[x];
                    waitQueue[x] = waitQueue[y];
                    waitQueue[y] = t;
                }
            }
        }
    }

...

PSA(优先级调度)

按照进程的优先级选择调度顺序

/**
 * 优先级调度算法
 * 就是将 SJF 算法中的排序，改为按照优先级排序
 */

/**
 * 改变部分
 * 对等待队列按进程优先级按从小到大排序
 */
for (int x=front, z=0; z < length; x=(x+1)%MAXLEN, z++){
    for (int y=x+1, q=0; q < length-x; y=(y+1)%MAXLEN, q++){
        if (processes.get(waitQueue[x]).getPriority() > processes.get(waitQueue[y]).getPriority()){
            int t = waitQueue[x];
            waitQueue[x] = waitQueue[y];
            waitQueue[y] = t;
        }
    }
}

RR (时间片轮转算法)

为 CPU 的执行设定一个时间片大小，每个进程轮询分配时间片，时间片结束后暂停运行加入等待队列

  /**
   * 时间片轮转算法
   * @param processes
   * @param round
   */
  public static void RR(List processes, int round){
      int count = processes.size();
      int time = 0;
      int[] waitQueue = new int[MAXLEN];
      int front = 0;int tail = 0;
      int running = 0;
      System.out.println("------------- RR 算法-------------");
      if (count <= 0){
          System.out.println("无可用进程");
          return;
      }

      while (count > 0){
          System.out.print("第 " + time + " 秒: ");
          for (int i = 0; i < processes.size(); i++){
              if (processes.get(i).isAlive() && processes.get(i).getInTime() == time){
                  System.out.print("进程 " + i + " 到来 ");
                  waitQueue[tail] = i;
                  tail = (tail+1) % MAXLEN;
              }
              if (processes.get(running).isAlive() && processes.get(running).getCount() == 0){
                  System.out.print("进程 " + running + " 结束运行 ");
                  processes.get(running).setAlive(false);
                  processes.get(running).setEndTime(time);
                  count--;
                  CPU = 1;
              }
          }
          if (CPU == 1 && front != tail){
              running = waitQueue[front];
              front = (front+1) % MAXLEN;
              System.out.print("进程 " + running + " 开始运行");
              int temp = processes.get(running).getCount();
              temp--;
              processes.get(running).setCount(temp);
              CPU = 0;
          } else if (CPU == 0){
              int temp = processes.get(running).getCount();
              temp--;
              processes.get(running).setCount(temp);
              if (time % round == 0){
                  System.out.print("进程 " + running + " 暂停运行");
                  waitQueue[tail] = running;
                  tail = (tail+1) % MAXLEN;
                  CPU = 1;
              }
          }
          time++;
          System.out.println();
      }
      System.out.println("---------------------------------");
      ShowResult(processes);
  }

算法运行结果显示

    /**
     * 输出时间统计结果
     * @param processes
     */
    public static void ShowResult(List processes){
        int averageTime = 0;
        for (int i = 0; i < processes.size(); i++){
            int inTime = processes.get(i).getInTime();
            int endTime = processes.get(i).getEndTime();
            averageTime += endTime-inTime;
            System.out.println("进程 " + i + " : 到来时间: " +
                    inTime + " 结束时间: " +
                    endTime + " 周转时间: " +
                    (endTime-inTime));
        }
        System.out.println("平均周转时间: " + averageTime / processes.size());
        System.out.println("---------------END--------------");
    }