Software Project Staffing

Question 1

Staffind for bug fixing

Answer 1

• Many bugs need to be fixed
• Productivity of developers might vary up to 1:20
• Ad hoc assignment likely is ineffective and inefficient
• Instead: Using optimized assignment as a basis for decision-making
• Empirical evaluation comparing baseline (manual) plan with optimized plans.

Question 2

Staffing Assumptions

Answer 2

1. Each developer fixes exactly one bug at a time.
2. Each bug is fixed by exactly one developer.
3. Once assigned to a bug, each developer is working continuously on this bug.
4. The productivity of a developer is inverse proportional to the effort requried to fix a bug (where the time is known)
5. The duration for fixing a bug is defined between closing and opening date as defined from the manual assignment divided by the ration in expertise btween the new versus manual assignment.

Question 3

Staffing for minimum duration

Answer 3

• Given:
  
  • List of features and their respective tasks
  • List of dependencies between tasks
  • Effort estimation for all the tasks of all the features
  • Pool of developers available
  • Skill (productivity) profile of developers
  • Availability intervals of developers
• Problem: Find a task schedule and an assignment of developers to tasks and a such that:
  
  • All tasks of all features are performed
  • All the dependencies between tasks need to be fulfilled
  • The total duration (make-span) is minimized.

Question 4

Productivity Assumptions

Answer 4

1. A developer having (average) productivity level 1 would need workload(q) days to perform task q.
2. A developer of productivity level prod(d,q) would need workload(q)/prod(d,q) days to perform the task.

Question 5

Staffing for Maximum Value

Answer 5

• Given:
  
  • List of features and their respective tasks
  • List of dependencies between tasks
  • Effort estimation of all the features
  • Pool of developers available
  • Availaility intervals of developers.
• Problem: Find a task schedule and an assignment of developers to tasks such that
  
  • All release due dates are fulfilled
  • All the dependencies between tasks need to be fulfilled
  • The total value of th release is maximized.