Time-Varying Priority Queuing Models for Human Dynamics arXiv:1109.5990

Hang-Hyun Jo with R. K. Pan and K. Kaski Department of Biomedical Engineering and Computational Science Aalto University, Finland

Queuing model for bursts [Barabasi, Nature ’05]

priority

the highest priority first + random selection

waiting time

small ¿

¿

large ¿

time

e-mail

P(¿) » ¿¡ ® ; ® = 1

Decreasing priority: memory/interest loss

priority

the highest priority first + random selection (p<1)

¿ (twitter)

(coffee break)

(research)

x0 x(t) = (t + 1)° x0 » U(0; 1) time

Waiting time distributions deterministic (p=1)

stochastic (p<1)

~ bimodal distribution [Wu et al., PNAS ’10]

Increasing priority: e.g. deadline effect the highest priority first + random selection (p<1) priority

(unbounded)

¿ x(t ¸ d) = x0 (bounded)

µ ¶¯ t x(t) = x0 d

t=d

time

Waiting time distributions deterministic (p=1)

stochastic (p<1)

d = 100

bounded:

unbounded:

P (¿) » ¿ ¡ 2

Empirical analysis: arXiv.org & Physical Review journals

Updating process of papers in arXiv.org ~ decreasing interests of authors with trembling hands? Reviewing process of papers in PR journals ~ unbounded deadline effect of reviewers?

Thank you!