Scientific Computing
sruthi Manalikar
ExperimentReproduction.pptxA computational consumer-driven market model: statistical properties and the underlying industry dynamics
Experiment reproduction
Joel Muhizi
CISC-601
Intro
There are several patterns and dynamics are observed about industry adoption and evolution
Existing studies mostly focus on the supply side
- S-shaped pattern in the adoption level of many industries
- Existence of an industry life cycle
- larger and/or longer-lasting firms present a lower growth rate
- existence of a turbulence phase before reaching industrial maturity
2
Goal
Analyze how these patterns and statistical regularities can emerge in computational demand-driven market model
Assumptions
Consumers are not utility maximizers
Consumers desires are driven by imitation and adaptation
Consumer heterogeneity, tolerance to change …
Data set and setup
Two sets represented in a characteristic space
Consumers
producers
Procedures
Iteration
Consumer side:
Local emulation
Innovation in consumption
Decision to purchase
Producer side:
Decision to enter
Decision to leave
Decision to adapt
Original Results
The model was able to reproduce the properties below
Smaller firms grow more quickly
The larger the firm, the less
The longer-lasting firms grow less
Longer-lasting firms offer less variability in their market share
Growth rates follow a distribution skewed to the right …
My results
Consumer side model
Producer side model
Reflections
Learnt new concepts such as clustering algorithms
Implemented a computational model that explains some real life observations
12
