Agent-based modeling for migration and modern slavery research: a systematic review

Abstract

This systematic review aims to synthesize how agent-based modeling (ABM) has been used in migration and modern slavery research and provide the basis to model development for social science researchers exploring the use of ABM. We searched five bibliographic databases using two terminology categories: (1) migration or modern slavery terminology; (2) complex system methods terminology. Two reviewers conducted independent article screening. Peer-reviewed articles presenting original migration or modern slavery ABMs were included. Data extraction included model development steps and model characteristics. The dataset was synthesized and compared across studies. We identified 28 articles for inclusion. Many of the ABMs tested theories and about half were based on empirical data. Model development varied considerably and reported methods were extremely opaque. Only five studies used a structured development framework. The most common model involved agents deciding whether and where to migrate and attempting migration. Climate change was a common exogenous scenario modeled. Most of the ABMs did not undergo any sensitivity analysis or validation.ABM has a greater capacity to account for heterogeneous and dynamic decision-making than more frequently applied methods in research on migration and modern slavery. However, there is still a paucity of studies adopting ABM methods. These reviewed ABMs highlight gaps in the reporting and implementing of model development. ABM is a promising technique to address many urgent and complex questions in research on migration and modern slavery to better support decision-makers, but addressing current methodological gaps is a critical first step.

The role of helplines in the anti-trafficking space: examining contacts to a major ‘modern slavery’ hotline

“…Second, anti-trafficking helplines are a recent phenomenon: even the apparently longest-running national helpline only opened in 2007 (Polaris, 2017). Third, the organised crime frame has long dominated anti-trafficking, and calls for a public health approach are only recently becoming more mainstream (McAlpine et al, 2021;.…”

A Bayesian Analysis of Migration Pathways using Chain Event Graphs of Agent Based Models

“…Despite their ability to plausibly model the transitions of an agent, many ABMs, both in migration research [13] and more broadly [9,5], have been described as opaque with many of the critical details needed to fully understand or replicate the models missing from publication, such as the lack of standardising model development. Some attempts, such as the ODD protocol, have been made to create a standardised structure for explaining ABMs [5], but there is still significant variance in how the protocol is used and the clarity it brings to ABMs.…”

“…Increasingly, Agent Based Models (ABMs) have been commonly used in contexts such as migration as they focus on the level of the individual and can be constructed by modelling the potential outcomes of successive events and decisionmaking [13,11]. In order to construct these models, a range of data sources, such as large structured demographic datasets or natural language narratives and theories have been used to inform deterministic and stochastic transitions within an ABM.…”

Operations research and analytics to combat human trafficking: A systematic review of academic literature

“…Owing to the breadth of domains that contribute to anti-HT research, a wealth of literature exists that has been well-documented in surveys over the years [ 10 , 11 , 17 – 21 ]. Existing reviews focus on social science, healthcare, and law enforcement approaches; whereas OR and Analytics have much to offer [ 22 ], no systematic review exists for the emerging landscapes of Operations Research (OR) and Analytics as applied to anti-HT.…”