Due to massive increase in Web service in order to provide users an improved service it is necessary to develop a dynamic approach based on multi constraint Quality-of-Service (QoS) driven web service composition model to recommend suitable services to intended users is a big challenge. QoS plays a prominent role in selecting a Web Service. QoS is mainly determined by several non-functional parameters like availability, reliability, robustness, integrity, accessibility, interoperability, accuracy and security. It is evidenced based on experimentation that the proposed Improved Rider Optimization scheme achieves near optimal solution where group of riders racing towards a target location and attains adaptability and scalability by making a more confident web service selection with QoS prediction.