Abstract Concerning computational physics, web services are conceived as mathematical units that are experienced in different systems that offer service composition. Due to the exponential growth of web services and their deployment on cloud platforms, quality of service parameters have now become an essential factor when searching for and selecting services that must satisfy specific non-functional requirements of a user application. A variety of service components are highly configurable and are dynamic scenarios because a significant number of services can meet these requirements. This work analyzes the systemic perspective of approaches for the selecting and searching of web services that have specifications of optimization strategies based on the configurable quality of service parameters with test scenarios in cloud environments that have a considerable number of services as input. The study shows that policies based on artificial intelligence and related areas are the ones with the most significant convergence, and the approaches analyzed to give a perspective of future work aimed at strategies based on automatic learning.