When applying MWAF, several expert users will be asked to give two parameters to each of the evaluated attributes: a weight to identify the importance of the attribute, and a rate to measure its strength or effectiveness.
The MWAF is an eight step process as defined below (see Figure 2):
The major step for implementing MWAF is to conduct a statistical experiment to evaluate the given products (see Figure 2).
Compatibility: MWAF is capable to conduct evaluation studies that are similar to many cases presented in the literature.
Structure: MWAF can be represented by an effective hierarchical structure, which derives its power from the principle of 'divide and conquer' that contributes to analyzing a complete taxonomy of evaluation attributes.
Scalability: MWAF is flexible to scaling up/down in order to expand or reduce its dimensions and/or attributes.
The functionalities of GUI facilitate configuration of the MWAF framework, collection of data comprising expert users' ratings, initiation of the statistical evaluation process, and displaying of the analysis results.
Administrator logs into the GPSE system in order to create expert users' accounts, setup the MWAF configuration framework, initiate the statistical analysis on the ratings data provided by expert users, and display results for the software evaluation problem.
UCD2--Setup and Configuration of MWAF: This functionality deals with the configuration of MWAF architecture for the given evaluation problem.
Hence, the system fulfils its technical goals in that a functional GPSE system based on MWAF is developed and meets the desired objectives.
Furthermore, the MWAF process is also being automated using a multi-agent system in which intelligent software agents are responsible for steps of the MWAF process.