Home > Research > Activities > Agent-based Classroom Lesson Simulation
View graph of relations

Agent-based Classroom Lesson Simulation

Activity: Talk or presentation typesPublic Lecture/ Debate/Seminar

5/09/2019

The goal of the research was to develop an agent-based model and simulation that adequately represented some of the dynamic and spatial complexities of secondary school classroom lessons, with interactions between autonomous students, their teacher and an optional teaching assistant. By implementing the conceptual model as a simulation, the aim was to study quantitatively the consequences of common classroom scenarios, to gain behavioural insights into what happens and how this affects overall class productivity, disruption, etc.
Such a simulation model could be a valuable lesson analytics tool, enabling teachers and educators to explore the consequences of interventions prior to implementing the changes in lessons. Potential scenarios might include providing or withdrawing a teaching assistant, changing seating arrangements, or trying alternative lesson plans.
During a 5-month case study, a lesson event recording tool was developed and used to record many types of student and teacher activities in 42 hours of lessons. These data have been used to calibrate and validate the model, lesson by lesson. Various metrics have been identified or constructed to characterise lessons, classes, students and teachers in order that empirical results and the results of simulations may be compared and evaluated.
The simulation software, coded in NetLogo (a popular open-source tool), is fully operational, providing animated lesson play/replay with rich statistical outputs for lesson analytics. The software will be demonstrated.

Event (Conference)

Title61st Annual Conference of the Operational Research Society
Abbreviated titleOR61
Date3/09/195/09/19
LocationUniversity of Kent
CityCanterbury
CountryUnited Kingdom
Degree of recognitionInternational event