Hybrid Modelling Stream

Talks:

Hybrid Simulation as a Means to Deal with COVID-19

Hybrid simulation (HS) has existed for as long as simulation itself. Over the years HS has shown many promising characteristics for modelling complex systems, however, it has never been used as commonly as single techniques. Whilst there have been many attempts to improve HS development, a lack of joint modelling skills seems to be the main barrier to the progress of HS. This may be due to the difficulty of training or merely the individual traits linked to each technique.

The advent of COVID-19 highlighted the need for a HS approach. However, the development of HS models has a very long lead time and cannot cope with the rapid impacts of COVID-19. In this talk, we demonstrate how a HS model could be built using publicly available single-method models, which allows for faster model development and also deals with the challenge of joint skills.

Presenters:

This stream features three 15 minute talks from distinguished speakers and a 15 minute Q&A session to finish.

Tillal Eldabi, 
University of Surrey

How and when to combine system dynamics and agent-based simulation for modelling infectious diseases

Simulation modelling has been used extensively in epidemiology to study the transmission dynamics of infectious diseases and the impacts of different interventions. Recently, a large and growing body of literature has adopted system dynamics (SD) or agent-based modelling (ABM), to investigate the spread of COVID-19 and assist policymakers in making decisions on the most effective strategies to control the pandemic.

Some problems may require modellers to look at different levels and dimensions of a system which is composed of interactive and interconnected constituents with dynamic behaviours. Hybrid simulation models that combine the advantages of SD and ABM can be useful in finding the best solution for such problems. Despite the growing interest in this approach, developing hybrid SD-AB models has been a challenging task as guidance on when and how they should be combined is scanty.

This may limit the use of hybrid models in addressing important questions on infection prevention and control. In this session, we will present several designs for combining SD and ABM, provide examples of epidemic hybrid models from literature where available, and suggest relevant questions for modelling infectious diseases that may benefit from hybrid modelling. We will also discuss the design of the hybrid model that we develop to investigate the impact of staff sharing upon the spread of COVID-19 within a network of care homes.

A Systematic Literature Review of Hybrid Simulation in the Construction Industry

Hybrid Simulation (HS) is the application of multiple simulation techniques, e.g. ABS, SD, DES, in the context of a single simulation study. While HS is becoming popular in industries such as healthcare and transportation and logistics, its potentials in the construction industry are yet to be fully explored. This is in addition to the clear disconnect between advances in ORMS in the OR community and the application of modelling in Construction Engineering community.

In this talk, we present a review of the literature that analyses existing research in HS for construction to bridge this gap. We attempt to categorise and synthesise the research in the area through lessons derived from inter-disciplinary introspection of similar methods but in different application contexts. In such way, the OR community would learn about the latest advances of HS in this sector, whilst the Construction Engineering community would benefit from advances in HS techniques.