Knowledge-based Planning Systems and HTN Planners

Knowledge-based Planning Systems are based on a philosophy of using whatever domain knowledge is available to solve the planning problem. These systemscan tackle complex domain models, and are characterized by the use of multiple types of domain knowledge and complex domain models to support their reasoning processes. This knowledge may include task and goal structures, various kinds of constraint, search control techniques, and interaction with humans when necessary to make use of their expertise.

HTN stands for 'hierarchical transition network'. Typically, a planner starts with an abstract network representing the task to be solved, and proceeds by expanding the abstract network into more detailed networks lower down in a bstract hierarchy, until the networks contain executable actions. This greatly reduces the search space of the planner by encoding knowledge on how to go about looking for a plan in a domain and also enables the user to control the type of solutions that are considered.

O-Plan From Austin Tate et al at AIAI, University of Edinburgh.

SHOP from Dana Nau et al at the University of Maryland.

SIPE-2 by David Wilkins at SRI International.

HyHTN From Donghong Liu at University of Huddersfield.

TALplanner developed by Patrick Doherty and Jonas Kvarnström at Linköping University. Uses a Temporal Action Logic which enables it to plan for incompletely specified dynamic worlds.

TLplan from Fahiem Bacchus and Michael Ady at the University of Toronto.

UMCP From Parallel Understanding Systems Group, Computer Science Dept., University of Maryland at College Park.

Print Page     PLANET Home