Aggregator
SC-IPM uses aggregators (eg,
prod
,
sum
)
to combine the results of subsidiary
Variables
(eg,
assim_eff
,
conc
,
grazing_rate
).
Constraint
SC-IPM uses sets of constraints
(eg,
always-together
,
atmost-one
,
exactly-one
,
necessary
)
to control the inclusion of
Generic Processes
in a
Model Structure.
Equation
SC-IPM uses algebraic and differential equations
to specify mathematical relationships
among
Generic Entities
and
Generic Processes.
See the
Equation page for details.
Entity Instance
Each entity instance
(in the current
Instance Library)
provides SC-IPM with:
- the instance name (eg,
aurelia
, nasutum
)
- the instance type (eg,
grazer
, producer
)
- parameter types (eg,
assim_eff
, attack_rate
)
- variable bindings (eg,
conc
, grazing_rate
)
Entity Role
Each
Generic Process may be used in
one or more entity roles (eg,
G
,
P
), each of which has
one or more
Entity Types (eg,
grazer
,
producer
).
Entity Type
Each
Generic Entity and
Entity Instance
has an entity type (eg,
grazer
,
producer
).
Fitness Score
Parameter Estimation generates a fitness score
for each
Model.
This is used to sort and filter the list of models that SC-IPM displays.
Generic Entity
SC-IPM uses generic entities (eg,
Grazer,
Producer) to describe and define
the general nature of entities in the resulting models.
Each generic entity has an
Entity Types
(eg,
grazer
,
producer
).
Generic Library
The generic library (eg,
pplib
) tells SC-IPM what kinds of
Generic Entities and
Generic Processes are available
and what
Constraints limit their inclusion
in a
Model Structure.
Generic Process
SC-IPM uses generic processes (eg,
Grazing Predation,
Logistic Growth) to describe and define
the general nature of processes in the resulting models.
Induction
IPM uses induction to assist in the creation and evaluation
of
Models and
Model Specifications.
Inductive reasoning, also known as induction or informally "bottom-up" logic,
is a kind of reasoning that constructs or evaluates general propositions
that are derived from specific examples.
-- Inductive reasoning (WP)
Inductive Process Modeling
Inductive Process Modeling (IPM) assists in the creation of
Models
that are both descriptive and explanatory.
Using assorted techniques
(eg,
Induction,
Parameter_Estimation),
it helps the modeler to generate, evaluate, and modify models.
SC-IPM, in particular,
creates candidate
Model Structures,
based on
Model Specifications
(
Constraints,
Entity Instances,
Generic Entities, and
Generic Processes).
See the
Introduction page
for a summary of IPM's motivation and approach.
Instance Library
The instance library tells SC-IPM which
Generic Library,
Parameters, and
Variables to use,
what
Entity Instances
(eg,
aurelia
,
nasutum
) to create, etc.
Model
SC-IPM generates and evaluates mathematical models,
based on
Model Specifications
and a set of
Training Data.
Model Specification
SC-IPM's model specifications consist of
Constraints,
Generic Entities,
Generic Processes,
Parameters, and
Variables.
Model Structure
SC-IPM generates sets of candidate Model Structures,
based on a
Model Specification.
Parameter
Generic Entities and
Generic Processes
may have associated Parameters (ie, parametric constants).
Each Parameter (eg,
attack_rate
,
assim_eff
,
gmax
)
is specified with
lower-bound
and
upper-bound
values.
This limits the range of numeric values considered
during
Parameter Estimation.
Parameter Estimation
SC-IPM performs Parameter Estimation on validated
Model Structures,
fitting their
Parameters to the training data.
Satisfiability
Each candidate
Model Structure
must pass a
Boolean satisfiability test
to ensure that it has a valid structure
and meets all specified
Constraints.
Training Data
Each
Instance_Library (eg,
pp-instances
)
specifies a list of input data files (eg,
pp-sim
).
This data is used to "train" (and sometimes test)
the
Parameter Estimation.
Variable
Generic Entities and
Generic Processes
may have associated Variables.
Each Variable is specified by an
Aggregator (eg,
prod
,
sum
) or an
Equation.