Software requirement analysis

Software Requirements Analysis: Implications for Agent-Based Systems

The reliance on agent-based technologies to streamline the development, fine-tuning and implementation of software is an area that is seeing significant gains using constraint-based modeling and heuristic programming (Kumar, Goyal, 2011) in addition to significant gains in Web Services technologies (Maamar, Mansoor, 2003). The intent of this analysis is to summarize and critique the peer-reviewed research summary provided in Software Requirement Analysis Enhancements By Prioritizing Requirement Attributes Using Rank Based Agents (Kumar, Goyal, 2011) which shows the results of using agents to streamline key processes in the Software Development Lifecycle (SDLC).

Summary and Critique

The use of intelligent agents to streamline and optimize the SDLC process has shown potential to accelerate the use of explicit knowledge in the development, testing and finalized of software code and associated modules. This approach to using intelligent agents combines the analysis of historical data, using heuristics to interpret, project and define parameters that are relevant to the specific phase of the SDLC module and the components associated with a given development effort (Kumar, Goyal, 2011). The design of this specific methodology concentrated on the ability of agents to store historical data and interpret patterns through the use of role-based taxonomy definitions (Kumar, Goyal, 2011). This allowed for the creation of a requirements environment definition with could also create it sown requirements checklist over time, ensuring intelligent interpretation and definition of operational parameters that aligned to domain requirements (Kumar, Goyal, 2011). This also ensured that the requirement entities within a given problem domain can be assessed, modeled and using the heuristics and data structures within a given intelligent agent, define specific actions and activities by role (from a taxonomy standpoint) or using the classical SDLC framework (Kumar, Goyal, 2011). The development of intelligent agents as Web Services parallels this development as it highlights the need for greater clarity in the areas of SDLC integration and support (Maamar, Mansoor, 2003). Where the Web Services approach to defining intelligent agents seeks to create a universal state engine of specific intelligence across the entire SDLC cycle, the use of intelligent agents looks to define domain heuristics (Kumar, Goyal, 2011) then project code configurability based on results.

The challenges of both Web Services-based and heuristically designed intelligent agents are to take into account the innately unquantifiable aspects of the context of agent-based development and make them align to project goals. On this aspect of agent-based SDLC performance, both approaches are limited in terms of their applicability and scalability. The reliance on heuristics can only go so far with the embedding of business, technical and organizational elements into the overall structure of an SDLC methodology (Kumar, Goyal, 2011). The reliance on an agent-based model fits well with the development of modules that are designed to align with these innately unquantifiable aspects of the context of an SDLC project, and further, the use of the completed application. Web Services is more utilitarian in its definition of functionality and its need to be pervasive and accessible as an inherent design criterion (Maamar, Mansoor, 2003). This utilitarian approach to defining Web Services is in contrast to the highly specified configured parameters of an agent-based approach to SDLC-oriented heuristics (Kumar, Goyal, 2011). While each has its unique strengths and must be selectively applied base don the business objective of the software being developed, each need to make significant gains in quality management and self-optimization before these technologies will be used more pervasively in enterprise software development.