Strategy updating rules and strategy distributions in dynamical multiagent systems

Area of Science:

• Complex Systems
• Evolutionary Game Theory
• Agent-Based Modeling

Background:

• The evolutionary minority game models strategy adaptation in competitive environments.
• Prior research highlighted intriguing dynamics for prize-to-fine ratios less than unity.

Purpose of the Study:

• To investigate the system dynamics of the evolutionary minority game with a specific strategy updating rule.
• To understand how parameters like prize-to-fine ratio and strategy update scale influence population behavior.

Main Methods:

• Simulated the evolutionary minority game using a strategy updating rule: p → p ± delta(p).
• Analyzed the impact of prize-to-fine ratio (R), length scale delta(p), and boundary conditions on strategy distribution.
• Examined temporal oscillations in the gene space.

Main Results:

• Strategy distribution is highly dependent on R, delta(p), and boundary conditions.
• Parameters R, delta(p), and boundary conditions dictate the amplitude and frequency of temporal oscillations.
• These oscillations are key determinants of the population's strategy distribution.

Conclusions:

• Agents using a coin-tossing strategy (p=1/2) exhibit the highest survival probability in the long term.
• This survival advantage holds true regardless of the delta(p) value or boundary conditions employed.