We propose a new approach to quantitative business models based on Computational Economics, which encodes the rules of a market and preferences of individual agents to derive the overall market state, by aggregating all agent choices. This approach provides a superior alternative to classical economics theory, to tackle common business questions such as estimating the impact of a price change on consumer demand.
Two Pillars of BIG's Computational Economics Approach
Our Agent-Based Models de-average to the individual agent and aggregate the behaviors of all agents to simulate a complex environment.
Satisfiability Modulo Theories (SMT)
Our proprietary SMT solving approach combines search with deduction - pruning very large search spaces quickly - to solve millions of utility optimization problems in seconds.
Elements of BIG's Computational Economics Models
Learn preference distributions
Use historical data and/or consumer research to learn distributions for individual utility functions.
Create many (millions) of AI agents with unique utility functions
Encode preferences and complex market trade-offs automatically through SMT formulas.
Simulate agent and market interactions
Allow agents to make decisions at different times (not simultaneously) and update market state after each agent's decision for more realistic market behavior.