Honeywell's Interval Constraint Engine's (ICE) representation of temporal
relationships, realistic task and resource modeling, a powerful temporal
analysis engine, and sophisticated search algorithms have been combined to
solve a variety of large, complex planning, scheduling, resource allocation,
and replanning problems.
Java programming is used as the development and delivery vehicle for these
technologies to facilitate rapid integration with existing systems and data
file formats. The web-based networking and presentation capabilities supported
by Java make it possible to quickly develop custom user interfaces and access
channels.
Constraint-Based Representation Technology:
Plan objectives and schedule event relationships are described as relative
numeric and temporal constraints. A graph of these relationships describes the
space of possible schedules and supports efficient detection and resolution of
infeasible task orderings or deadlines.
Task and Resource Modeling Technology:
Tasks and resources are modeled in detail to ensure that all necessary
resources will be available for task execution. A rich set of resource classes
is supported, allowing accurate schedules to be generated for a broad range of
applications.
Temporal Analysis Technology:
The Interval Constraint Engine (ICE) efficiently computes the implications
of each planning and scheduling decision, providing rapid plan evaluation,
initial schedule generation, and nearly immediate replanning when disruptions
occur. On a typical workstation the ICE is capable of handling more than 25,000
tasks and 150,000 constraints in from 1 to 10 seconds.
Search Algorithm Technology:
Basic search algorithms such as depth first, back jumping, and dynamic
backtracking are used to conduct an initial search of the planning and
scheduling solution space or for assigning resources to scheduled tasks. Once
the nature of the solution space is understood, specific algorithms defining
appropriate planning or resource allocation policies can be rapidly implemented
and integrated for a custom solution.
These planning and scheduling technologies have been successfully applied to
such diverse applications as: space shuttle crew workloads, optimization of
demand flow manufacturing, 777 avionics processing, satellite image data
processing and archiving, batch factory manufacturing, aircraft repair and
maintenance, petroleum blending, and crude oil processing.
