A Minimal Agent-Based Model of Online Dating Markets (1.1.0)
A minimal agent-based model of a heterosexual swipe-based dating market. Agents carry a latent attractiveness coordinate, perception biases, gender-asymmetric downward acceptance thresholds, and exogenous long-term/short-term mating strategies drawn from survey shares. Each period, singles sample opposite-sex profiles, pursue and accept via mutual-consent logistic rules, and accumulate utility from matches (with diminishing returns in partner count), loneliness, rejection, and strategy-mismatch penalties.
Parameters use a three-tier governance scheme: empirically anchored (Tier A), structurally grounded (Tier B), and calibrated assumptions (Tier C). Only three primary swipe moments are calibrated; a further fifteen platform statistics are emergent checks scored by calibration- and literature-distance (18-moment scorecard). The model is analysed for welfare inequality, including a negative-utility tail dominated by never-matched men.
The model is documented following the ODD protocol (see model/spec.md). Associated manuscript submitted to the Journal of Artificial Societies and Social Simulation (JASSS), 2026.
Release Notes
Associated Publications
Manuscript under review at the Journal of Artificial Societies and Social Simulation (JASSS), 2026. Author details withheld for blind peer review.
This submission has been flagged as spam and is awaiting moderator approval.
Believe this is a mistake? Please
contact us!
A Minimal Agent-Based Model of Online Dating Markets 1.1.0
Submitted by
paved-bio
Published Aug 24, 2026
Last modified Aug 24, 2026
A minimal agent-based model of a heterosexual swipe-based dating market. Agents carry a latent attractiveness coordinate, perception biases, gender-asymmetric downward acceptance thresholds, and exogenous long-term/short-term mating strategies drawn from survey shares. Each period, singles sample opposite-sex profiles, pursue and accept via mutual-consent logistic rules, and accumulate utility from matches (with diminishing returns in partner count), loneliness, rejection, and strategy-mismatch penalties.
Parameters use a three-tier governance scheme: empirically anchored (Tier A), structurally grounded (Tier B), and calibrated assumptions (Tier C). Only three primary swipe moments are calibrated; a further fifteen platform statistics are emergent checks scored by calibration- and literature-distance (18-moment scorecard). The model is analysed for welfare inequality, including a negative-utility tail dominated by never-matched men.
The model is documented following the ODD protocol (see model/spec.md). Associated manuscript submitted to the Journal of Artificial Societies and Social Simulation (JASSS), 2026.