Quantification of deviations from rationality with heavy tails in human dynamics

Area of Science:

• Complex systems
• Network science
• Computational social science

Background:

• Human task dynamics influence technological, economic, and social systems.
• Queueing theory suggests power-law distributions for task waiting times.
• The universality and limits of power-law behavior in task completion are not fully understood.

Purpose of the Study:

• To investigate the regimes of task completion times for internet users.
• To identify deviations from the pure power-law model in real-world data.
• To understand how user behavior and system load affect task fulfillment.

Main Methods:

• Analysis of anonymized World Wide Web data from Google.
• Characterization of task execution times after message reception.
• Identification of different statistical regimes based on user utilization and response strength.

Main Results:

• The pure power-law distribution is not universally applicable to task completion times.
• Identified an exponential regime for tasks performed with minimal delay.
• Observed a crossover to an asymptotic plateau, indicating some tasks are never completed.

Conclusions:

• Task completion dynamics are more complex than previously modeled by simple queueing theory.
• User time utilization and response to stimuli create distinct performance regimes.
• Findings have implications for understanding online system performance and user behavior.