Thèse préparée par Edison CHUNG

Title :  Interoperability in Decentralized Goal Directed Building Management System

Beginning of the­sis : 2020‐​01‐​01
End of the­sis : 2022‐12‐31

Abstract : As the BIM domain under­goes a digi­tal trans­for­ma­tion, the deployed ser­vices must be able to react and adapt to the conti­nuous­ly chan­ging envi­ron­ment. Given the com­plexi­ty and time res­ponse constraints, the decision‐​making is being dele­ga­ted to buil­ding mana­ge­ment sys­tems, thus trans­for­ming BIM sys­tems from being pure­ly data‐​driven to beco­ming goal‐​directed. MASs are pro­mi­sing solu­tions to sup­port this trans­for­ma­tion, as agents will be able to control one or more ser­vices and to coor­di­nate and regu­late themselves.

These BIM sys­tems are beco­ming increa­sin­gly com­plex, dyna­mic, and hete­ro­ge­neous. However, the domain spe­cia­li­zed, indi­vi­dual sys­tems must remain inter­ope­rable to have a flexible and decen­tra­li­zed deci­sion sys­tem. This inter­ope­ra­bi­li­ty implies coping with data, pro­to­col, and ser­vice hete­ro­ge­nei­ties. In this regard, Semantic Web tech­no­lo­gies can help fos­ter uni­for­mi­ty of data for­mats, as well as modu­la­ri­za­tion and reuse of spe­ci­fi­ca­tions (onto­lo­gies).

While both MAS and Semantic Web tech­no­lo­gies have been used inde­pen­dent­ly in solu­tions within the buil­ding domain, there doesn’t exist a ful­ly inte­gra­ted infra­struc­ture where all these pro­po­sals are inter­ope­rable. Thus, the main objec­tive of this PhD is to pro­pose a solu­tion capable of leve­ra­ging the advan­tages of both MAS and Semantic Web tech­no­lo­gies, by defi­ning models and tech­no­lo­gies for a ful­ly decen­tra­li­zed goal‐​directed buil­ding mana­ge­ment system.

Keywords : Semantic Web Technologies, Multi‐​Agent Systems, Web of Things, Thing Description, Building Management System

Thesis defense :  2022‐12‐31

Supervisor :

Partners or/​and fun­ders : ENGIE LAB Laboratoire CSAI