2026-10-08T15:23:48Z https://drops.dagstuhl.de/oai

oai:drops-oai.dagstuhl.de:23890 2025-11-12T12:11:03Z ddc:004 open_access

From Prediction to Action: A Constraint-Based Approach to Predictive Policing Mechqrane, Younes Elabbassi, Ismail Inductive Constraint Programming (ICON) Loop Next Frame Prediction PredRNN++ Crime prevention in urban environments demands both accurate crime forecasting and the efficient deployment of limited law enforcement resources. In this paper, we present an integrated framework that combines a machine learning module (i.e. PredRNN++ [Wang et al., 2018]) for spatiotemporal crime prediction with a constraint programming module for patrol route optimization. Our approach operates within the ICON loop framework [Bessiere et al., 2017], facilitating iterative refinement of predictions and immediate adaptation of patrol strategies. We validate our method using the City of Chicago Crime Dataset. Experimental results show that routes informed by crime predictions significantly outperform strategies relying solely on historical patterns or operational constraints. These findings illustrate how coupling predictive analytics with constraint programming can substantially enhance resource allocation and overall crime deterrence. Schloss Dagstuhl – Leibniz-Zentrum für Informatik Younes Mechqrane and Ismail Elabbassi 2025 Is Part Of LIPIcs, Volume 340, 31st International Conference on Principles and Practice of Constraint Programming (CP 2025) InProceedings Text doc-type:ResearchArticle publishedVersion application/pdf doi:10.4230/LIPIcs.CP.2025.29 urn:nbn:de:0030-drops-238902 https://drops.dagstuhl.de/entities/document/10.4230/LIPIcs.CP.2025.29 eng https://creativecommons.org/licenses/by/4.0/legalcode