Experimental (raw) Data of Statistical Comparison Between Formal and Simulated Models’ Outcomes for CIBI vs. CVP General Problem

All the numerical values of the experiment/simulation process are provided in a a single .xlsx file, organized per simulation state on 8 separate sheets

Further analysis and statistical techniques can be easily implemented using the provided (raw) data to extract or reveal alternate evidence and reach new conclusions

In each sheet, columns are related to:

A: number of samples instance which corresponds to a specific system

B,C, D, E: values of distinct design attributes, scenario probability, and developers' experience level of the specific instance/system

F to Y: values of the Simulation Model (SM) for different 'lamda' values (number of scenarios' applications 5,10, .. 95, 100), concerning CIBI design combination

Z to AS: values of the Simulation Model (SM) for different 'lamda' values (number of scenarios' applications 5,10, .. 95, 100), concerning CVP design combination

AT to BM: values of the Simulation Model (SM) for different 'lamda' values (number of scenarios' applications 5,10, .. 95, 100), concerning the differnce of CVP-CIBI design combination

BN to CG: values of the Formal Model (FM) for different 'lamda' values (number of scenarios' applications 5,10, .. 95, 100), concerning CIBI design combination

CH to DA: values of the Formal Model (FM) for different 'lamda' values (number of scenarios' applications 5,10, .. 95, 100), concerning CVP design combination

DB to DU: values of the Formal Model (FM) for different 'lamda' values (number of scenarios' applications 5,10, .. 95, 100), concerning the differnce of CVP-CIBI design combination

DV to EO: values of error rate assessments (indicated by different signs among FM and SM related values concerning the difference of CVP-CIBI) after repeated executions of the SM for the same instance/system

Each sheet is related to different simulation state:

1: Deterministic results (no stochastic or random behavior)

2: Engagement of random sequences of scenarios' implementation

3: Engagement of alternate maintenance scenarios

4: Engagement of actual/fluctuating intervention's code size

5: Engagement of fixed developers' experience level and codes aging issues

6. Engagement of increased developers' experience level

7. Engagement of normal distributed developers' experience level

8. Engagement of left-skewed normal distributed developers' experience level

Scope of experiment:

-Analyze the FM of the CIBI vs CVP example problem

-for the purpose of validation, and thus the evaluation of the introduced theory and Modeling Methodology

-with respect to their reliability in terms of accurate effort predictions and thus valid decisions/selections among implementations alternatives

-from the point of view of the software engineer (researcher)

-in the context of the researcher effort to develop and execute a SM that imitates s/w actual maintenance process and returns effort computations, for a random sample of systems (software design attributes and scenario probabilities), in a computer lab, while the computed effort outcomes are compared to FM effort predictions.

Sample size: 1000 randomly selected observations (design attributes and developers experience level)

Output: (Direct) Formal Model predictions, Simulation Model computations of effort/size assessments, and Error rate assessments for each experiment scenario of the sample and for eight different SM states

Output size: 1,120,000 values