↓ Skip to main content
Milav
  1. Resources/
  2. Study Materials/
  3. Information & Communication Technology Engineering/
  4. ICT Semester 5/
  5. Software Engineering (4353202)/

3 mins· ·
Milav Dabgar
Author
Milav Dabgar
Experienced lecturer in the electrical and electronic manufacturing industry. Skilled in Embedded Systems, Image Processing, Data Science, MATLAB, Python, STM32. Strong education professional with a Master’s degree in Communication Systems Engineering from L.D. College of Engineering - Ahmedabad.
Lecture 29: Project Estimation Techniques - Analytical

Lecture 29: Project Estimation Techniques - Analytical

Unit 4: Software Project Management (4353202)

Lecture Agenda

  • Recap of Heuristic Estimation
  • What is Analytical Estimation?
  • Technique 1: Putnam's Norden/Rayleigh Curve
  • Technique 2: Function Point Analysis (Revisited)
  • Technique 3: COCOMO (Revisited as Analytical)
  • Advantages of Analytical Techniques
  • Disadvantages of Analytical Techniques
  • Key Takeaways

Recap of Heuristic Estimation

Heuristic estimation relies on expert judgment, analogy, and consensus (Delphi) for quick, experience-based estimates. They are flexible but can be subjective.

What is Analytical Estimation?

Analytical estimation techniques are based on mathematical formulas and models derived from the underlying structure of the software development process. They attempt to predict effort, cost, and schedule by analyzing the relationships between various project parameters.

These techniques are more formal and objective than heuristic methods.

Technique 1: Putnam's Norden/Rayleigh Curve

This model, developed by Lawrence Putnam, is based on the observation that software development effort and staffing levels follow a Rayleigh distribution curve over time.

  • Concept: Effort and staff build up slowly, peak, and then decline as the project nears completion.
  • Formula: Relates effort, time, and productivity parameters.
  • Advantages: Provides insights into staffing levels over time, useful for long-term planning.
  • Disadvantages: Assumes a fixed development process, less flexible for iterative models, and parameters can be hard to determine accurately.

Technique 2: Function Point Analysis (Revisited)

While Function Points (FP) are a size metric, they are also a core component of analytical estimation when used with productivity rates.

  • Concept: Once FP are calculated, historical productivity data (e.g., person-hours per FP) can be used to estimate effort.
  • Formula: Effort = Function Points / Productivity Rate
  • Advantages: Language-independent, focuses on functionality, and can be estimated early.
  • Disadvantages: Subjectivity in FP counting, and accurate productivity data is essential.

Technique 3: COCOMO (Revisited as Analytical)

The COCOMO model, particularly its Intermediate and Detailed versions, can be considered an analytical model because it uses mathematical formulas and adjusts them based on various project attributes (cost drivers).

  • Concept: Effort = a * (KLOC)^b * EAF (Effort Adjustment Factor).
  • Advantages: Widely accepted, provides a structured approach, and accounts for various project complexities.
  • Disadvantages: Relies on accurate KLOC estimation, and the EAF values can be subjective.

Advantages of Analytical Techniques

  • Objectivity: Based on mathematical models, reducing subjectivity and bias.
  • Repeatability: Given the same inputs, the models produce the same estimates.
  • Transparency: The underlying assumptions and calculations are explicit.
  • Predictive Power: Can provide more accurate estimates for complex projects, especially when calibrated with historical data.

Disadvantages of Analytical Techniques

  • Data Dependency: Require significant historical data for calibration and validation.
  • Complexity: Can be complex to understand and apply, requiring specialized knowledge.
  • Rigidity: May not adapt well to unique or rapidly changing project environments.
  • Initial Effort: Setting up and calibrating these models can be time-consuming.

Key Takeaways

  • **Analytical estimation** uses **mathematical models** for more objective and repeatable estimates.
  • Techniques include **Putnam's model, Function Point Analysis (with productivity), and COCOMO**.
  • They offer **higher accuracy and transparency** but require **more data and expertise**.
  • Often combined with heuristic methods for a balanced approach.

Next Lecture

Topic: Project Scheduling - Gantt Chart

Q & A

Questions & Discussion