策略模式
Define a family of algorithms,encapsulate each one,and make them interchangeable.
(定义一组算法,将每个算法都封装起来,并且使它们之间可以互换。)
通用类图
封装类，承上启下，屏蔽高层模块对策略、算法的直接访问。与抽象策略类是聚合关系
抽象策略类，策略、算法的抽象，定义必须具有的方法和属性
具体策略类，实现抽象的操作，含有具体的方法，继承抽象策略类
通用源码
//抽象策略类
class Strategy {
public:
//策略模式的运算法则
virtual void doSomething() = 0;
};
//具体策略类
class ConcreteStrategy1 : public Strategy {
public:
virtual void doSomething() {
//具体策略1的运算法则
}
};
class ConcreteStrategy2 : public Strategy {
public:
virtual void doSomething() {
//具体策略2的运算法则
}
};
//封装类
class Context {
private:
Strategy* strategy = NULL;
public:
Context(Strategy *s): strategy(s) { }
void doAnything() {
strategy->doSomething();
}
};
优点
算法可以自由切换
避免使用多重条件判断
扩展性良好
缺点
策略类数量增多
所有的策略类都需要对外暴露
使用场景
多个类只有在算法或行为上稍有不同的场景
算法需要自由切换的场景
需要屏蔽算法规则的场景
示例代码
#include <iostream>
#include <string>
using namespace std;
class IStrategy {
public:
virtual void operate() = 0;
};
class BackDoor : public IStrategy {
public:
virtual void operate() {
cout << "find helper, add pressure" << endl;
}
};
class GivenGreenLight : public IStrategy {
public:
virtual void operate() {
cout << "open green light, let out" << endl;
}
};
class BlockEnemy : public IStrategy {
public:
virtual void operate() {
cout << "lady Sun in back" << endl;
}
};
class Context {
private:
IStrategy *strategy = NULL;
public:
Context(IStrategy *is): strategy(is) { }
void operate() {
strategy->operate();
}
};
int main() {
Context *context = NULL;
context = new Context(new BackDoor());
context->operate();
context = new Context(new GivenGreenLight());
context->operate();
context = new Context(new BlockEnemy());
context->operate();
}
