Skip to Main content Skip to Navigation
New interface
Conference papers

Extraction of Agent Groups with Similar Behaviour Based on Agent Profiles

Abstract : This paper deals with the log files suitable to extract valuable information about agents and their behaviour from agent-based simulation in a model of virtual company. Such information, presented in a transparent way, can be used as a support for simulation verification to achieve the suitable design of the proposed system. Hence, based on the similar behaviour (represented by extracted sequences) of agents, we are able to construct models which explain certain aspects of agent behaviour. Moreover, we can extract agent profiles based on behaviour and find latent ties between different agent groups with similar behaviours. The paper extends the results of our previous works about sequence extraction and comparison. The approach for agent network construction based on agent profiles is described. Two different methods were used for construction of agent network. One method uses cosine similarity and graph partitioning and the second self organization maps and Euclidean similarity for agent relations. Each of these methods has its advantages and disadvantages which are summarized in the paper and presented in the form of the visualization of relations between agents.
Complete list of metadata

Cited literature [20 references]  Display  Hide  Download
Contributor : Hal Ifip Connect in order to contact the contributor
Submitted on : Monday, March 27, 2017 - 11:01:41 AM
Last modification on : Saturday, June 1, 2019 - 11:34:02 AM
Long-term archiving on: : Wednesday, June 28, 2017 - 1:13:29 PM


Files produced by the author(s)


Distributed under a Creative Commons Attribution 4.0 International License



Kateřina Slaninová, Jan Martinovič, Roman Šperka, Pavla Dráždilová. Extraction of Agent Groups with Similar Behaviour Based on Agent Profiles. 12th International Conference on Information Systems and Industrial Management (CISIM), Sep 2013, Krakow, Poland. pp.348-357, ⟨10.1007/978-3-642-40925-7_32⟩. ⟨hal-01496081⟩



Record views


Files downloads