Model Analysis Workshop (2011)

Informações do Evento

November 14-15, Brasília – DF, Brazil

Coordinator: Paulo Gonçalves

Prof. Paulo Gonçalves, PhD, is associate professor at the Institute of Management at the University of Lugano and research affiliate in the System Dynamics Group at the MIT Sloan School of Management. He is also the academic director of the Master in Advanced Studies in Humanitarian Logistics and Management (MASHLM), a part time master program intended for professionals in the humanitarian sector. He is interested in managerial decision making in traditional operational settings.

This workshop aims to provide a practical experience in developing a System Dynamics model in an interactively way. Participants will have a unique opportunity to learn and exchange knowledge and information and interact with one of the leading experts in System Dynamics, experiencing the major problems and difficulties in developing a complex model of System Dynamics and how to overcome them.

Day 1: November 14

Session #1 – The Market Growth Model (MGM)

Purpose: Introduce the MGM, review CLD, SFs, and Vensim syntax

Intro to MGM problem and model

• Feedback processes and causal loop diagrams in MGM
• Stock-and-flow structure in the MGM
• Implementing the Market Growth Model

Session #2 – Partial and full model analysis (MGM)

Purpose: Partial and Full model analysis

• MGM – Dynamic hypotheses:
  • Sales growth, Capacity Acquisition, sales effectiveness.
• MGM – Intended Rationality:
  • Indicated salesmen, delivery delays, capacity investment.
• Extreme condition tests:
  • (High–low) backlog, (Many–few) salesmen, (High–low) capacity.
• MGM – Full model analysis: Causal tracing

Day 2: November 15

Session #1 – Advanced methods: Sensitivity analysis (MGM)

Purpose: Univariate and multivariate sensitivity analysis

• Univariate sensitivity analysis
• Single parameter uncertainty: random variables and distributions
• Sensitivity tests: Salesman salary, delivery delay bias, capacity ordering delay
• Analyzing sensitivity results
  • Probability distributions, confidence intervals, graphs
• Multivariate sensitivity analysis
  • Monte Carlo simulation
  • Latin hypercube

Session #2 – Advanced methods: Optimization and linearization (MGM)

Purpose: Introduce/implement optimization, linearization, and calibration

• Optimization: Objective maximization/minimization
  • Designing an objective function
    • Multi-criteria decisions
    • Time discounting
• Linearizing the MGM
  • MGM’s systems of equations
  • Eigen-analysis MGM
• Calibration to data (time permitting)
  • Theil Statistics

Anúncio