CellML Discussion List

Text archives Help


[cellml-discussion] Seminar on ModML (including how it relates to CellML) today, 4-5 PM, 5th floor seminar room, 70 Symonds St


Chronological Thread 
  • From: ak.miller at auckland.ac.nz (Andrew Miller)
  • Subject: [cellml-discussion] Seminar on ModML (including how it relates to CellML) today, 4-5 PM, 5th floor seminar room, 70 Symonds St
  • Date: Tue, 17 May 2011 13:16:52 +1200

Hi all,

I'm giving a seminar today on ModML; I plan to discus how ModML fits
into the picture along with imperative and declarative modelling
languages (including CellML).

Sorry for the spam for those of you not in Auckland.

Best wishes,
Andrew

Details:
Venue: 5th Floor Seminar Room, 70 Symonds St
Abstract:
Mathematical modelling is central to the understanding of biological
systems and how they interact. Models of differential-algebraic
equations are an important sub-class of such models.

The current methods for representing systems of differential algebraic
equations all tend to be based either on imperative programming, mixing
the numerical algorithm with the model specification, thereby limiting
the reusability of the model, or on a declarative representation of the
mathematics. Declarative representations have the benefit that they
separate the numerical algorithm from the system, and can be used for
different purposes by different software programs. In declarative
languages like CellML, models are written in a generic
(non-domain-specific) form as equations; patterns in the equations such
as the relationship between fluxes and rates are repeated in the model,
and the biological meaning of the model needs to be annotated separately
in parallel to the equations. Some declarative languages, like SBML,
provide facilities for specific modelling domains (such as representing
reactions), but to add new domains, the language itself needs to change.

In this seminar, I present an alternative approach, in which models are
represented as a functional computation that produces a data-structure
representing differential-algebraic equations. Because the data
structure can be produced by functional transformations, modellers can
import an existing domain-specific transformation function and work on
the model in a domain-specific language; new domain-specific languages
can be created by the modeller without changing the language, and models
from different domains of application can be inter-connected.



  • [cellml-discussion] Seminar on ModML (including how it relates to CellML) today, 4-5 PM, 5th floor seminar room, 70 Symonds St, Andrew Miller, 05/17/2011

Archive powered by MHonArc 2.6.18.

Top of page