策略模式

策略模式定义了一系列的算法，并将每一个算法封装起来，而且使它们还可以相互替换。策略模式让算法独立于使用它的客户而独立变化。

策略模式里涉及到以下几个角色：

1.Context类，表示应用场景，包含一个策略类的基类指针

2.Strategy类，抽象类，定义了策略类的接口

3.ConcreteStrategy，具体策略类，一般有多个，实现了Strategy定义的接口

策略模式的UML图：

策略模式中每个算法都被封装成了类，可以通过自己的接口单独测试，简化了单元测试，同时避免了程序中过多的条件转移语句，简化了系统并使整个系统易于扩展，遵守了高内聚低耦合的设计原则。但同时，为每一个策略都产生一个类的做法增加了系统中类的个数，并且客户端必须了解这些算法的区别，算法类的选择由客户端负责。

C++实现：

namespace StrategyPattern{class Strategy//定义算法类接口{public:virtual ~Strategy(){}virtual void algorithmInterface() = 0;};class Context//算法类的应用场景{public:Context(){}Context(Strategy *pStrategy) :mPStrategy(pStrategy){}void setStrategy(Strategy* pStrategy){mPStrategy = pStrategy;}void ContextInterface(){mPStrategy->algorithmInterface();}private:Strategy *mPStrategy;};class ConcreteStrategy1 :public Strategy//具体策略类1{public:~ConcreteStrategy1(){}void algorithmInterface(){cout << "using ConcreteStrategy1" << endl;}};class ConcreteStrategy2 :public Strategy//具体策略类2{~ConcreteStrategy2(){}void algorithmInterface(){cout << "using ConcreteStrategy2" << endl;}};class ConcreteStrategy3 :public Strategy//具体策略类3{~ConcreteStrategy3(){}void algorithmInterface(){cout << "using ConcreteStrategy3" << endl;}};void test(){Context ctt1;Strategy *s1 = new ConcreteStrategy1();Strategy *s2 = new ConcreteStrategy2();Strategy *s3 = new ConcreteStrategy3();ctt1.setStrategy(s1);ctt1.ContextInterface();ctt1.setStrategy(s2);ctt1.ContextInterface();ctt1.setStrategy(s3);ctt1.ContextInterface();delete s1;delete s2;delete s3;}}

java代码：

public class Context {static void process(Strategy sttg ){sttg.algorithmsInterface();}public static void main(String[] args) {// TODO Auto-generated method stubprocess(new ConcreteStrategy1());process(new ConcreteStrategy2());}}interface Strategy{void algorithmsInterface();}class ConcreteStrategy1 implements Strategy{public void algorithmsInterface(){System.out.println("ConcreteStrategy1");}}class ConcreteStrategy2 implements Strategy{public void algorithmsInterface(){System.out.println("ConcreteStrategy2");}}