No associated github account.
Eric is a Research Fellow in the Complexity programme at the MRC/CSO Social and Public Health Unit at the University of Glasgow, working on agent-based simulation approaches to complex public health issues. Prior to this he was a Research Lecturer/Senior Lecturer in Artificial Intelligence and Interactive Systems in the School of Computing at Teesside University. Before working at Teesside, he worked on the CLC Project at the University of Southampton, a multidisciplinary project which focuses on the application of complexity science approaches to the social science domain.
Eric received a BA with Honours in Psychology from Pennsylvania State University, and a PhD from the School of Computing at the University of Leeds. After his PhD, he worked as a JSPS Postdoctoral Research Fellow at the University of Tokyo, conducting research in computer simulation and robotics.
A reimplementation of the Wedding Ring model by Francesco Billari. We investigate partnership formation in an agent-based framework, and combine this with statistical demographic projections using real empirical data.
This model is an agent-based simulation written in Python 2.7, which simulates the cost of social care in an ageing UK population. The simulation incorporates processes of population change which affect the demand for and supply of social care, including health status, partnership formation, fertility and mortality. Fertility and mortality rates are drawn from UK population data, then projected forward to 2050 using the methods developed by Lee and Carter 1992.
The model demonstrates that rising life expectancy combined with lower birthrates leads to growing social care costs across the population. More surprisingly, the model shows that the oft-proposed intervention of raising the retirement age has limited utility; some reductions in costs are attained initially, but these reductions taper off beyond age 70. Subsequent work has enhanced and extended this model by adding more detail to agent behaviours and familial relationships.
The version of the model provided here produces outputs in a format compatible with the GEM-SA uncertainty quantification software by Kennedy and O’Hagan. This allows sensitivity analyses to be performed using Gaussian Process Emulation.
This simulation investigates the provision and receipt of social care in a simulated UK population. Agents may decide to provide care, either informally or by paying for private carers, when a member of their kinship network exhibits care need. Care-giving decisions are informed by agents’ health status, employment status, closeness of their relationship to the affected agent, and geographical proximity. Agents may undergo various life-course transitions, including partnership formation and dissolution, migrating domestically, having children, and changing jobs. The results indicate that the model produces realistic patterns of care provision and receipt, despite the relative paucity of empirical data to inform the model.
The purpose of this model is the simulation of social care provision in the UK, in which individual agents can decide to provide informal care, or pay for private care, for their loved ones. Agents base these decisions on factors including their own health, employment status, financial resources, relationship to the individual in need and geographical location. The model simulates care provision as a negotiation process conducted between agents across their kinship networks, with agents with stronger familial relationships to the recipient being more likely to attempt to allocate time to care provision. The model also simulates demographic change, the impact of socioeconomic status, and allows agents to relocate and change jobs or reduce working hours in order to provide care.
Despite the relative lack of empirical data in this model, the model is able to reproduce plausible patterns of social care provision. The inclusion of detailed economic and behavioural mechanisms allows this model to serve as a useful policy development tool; complex behavioural interventions can be implemented in simulation and tested on a virtual population before applying them in real-world contexts.