Creating satisficing planners with deep learning
View Patent ↗Some embodiments of the present invention are directed to a method of choosing the components for a satisficing planner using machine learning (ML) (for example, deep learning (DL)). Some embodiments of the present invention are directed to choosing search algorithm component(s) for a satisficing planner using ML (for example, DL). Some embodiments of the present invention are directed to choosing search refinement components (that is, search boosting component(s) and/or search pruning component(s)) for a satisficing planner using ML (for example, DL).
1. A computer-implemented method comprising:
receiving a planner application data set that includes information indicative of contexts in which a newly-designed satisficing planner is expected to be applied;
receiving a machine learning algorithm that has been trained to design satisficing planners; and
applying the machine learning algorithm to the planner application data set to obtain a newly-designed satisficing planner design, with the newly-designed satisficing planner design including the information indicative of a plurality of satisficing planner components to be included in the newly-designed satisficing planner design and an order in which the satisficing components are to be run, wherein the applying of the machine learning algorithm includes the following operations:
for each given satisficing planner component of the plurality of satisficing planner components, determining a component type from a plurality of component types for the given satisficing planner component, wherein the plurality of component types are selected from a list consisting of heuristic function types, search algorithm types, and refinement algorithm types, and wherein the refinement algorithm types comprise a search boosting component subtype and a search pruning component; and
for each given satisficing planner component of the plurality of satisficing planner components, determining a component candidate, from among a plurality of component candidates, of the determined component type of the given satisficing planner component.
2. The method of claim 1 , wherein the machine learning algorithm is a deep learning algorithm that includes a plurality of layers.
3. The method of claim 2 , wherein the plurality of layers includes a plurality of hidden layers.
4. The method of claim 1 , wherein:
the planner application data set includes a previously used satisficing planner design; and
the application of the machine learning algorithm uses the previously used satisficing planner as a basis for the newly-designed satisficing planner design.
5. The method of claim 1 , wherein the newly-designed satisficing planner design includes:
at least one satisficing planner component of the heuristic function type;
at least one satisficing planner component of the search algorithm type; and
at least one satisficing planner component of the search refinement type.
6. A computer system comprising:
one or more processors, one or more computer-readable memories, one or more computer-readable tangible storage medium, and program instructions stored on at least one of the one or more tangible storage medium for execution by at least one of the one or more processors via at least one of the one or more memories, wherein the computer system is capable of performing a method comprising:
receiving a planner application data set that includes information indicative of contexts in which a newly-designed satisficing planner is expected to be applied;
receiving a machine learning algorithm that has been trained to design satisficing planners; and
applying the machine learning algorithm to the planner application data set to obtain a newly-designed satisficing planner design, with the newly-designed satisficing planner design including the information indicative of a plurality of satisficing planner components to be included in the newly-designed satisficing planner design and an order in which the satisficing components are to be run, wherein the applying of the machine learning algorithm includes the following operations:
for each given satisficing planner component of the plurality of satisficing planner components, determining a component type from a plurality of component types for the given satisficing planner component, wherein the plurality of component types are selected from a list consisting of heuristic function types, search algorithm types, and refinement algorithm types, and wherein the refinement algorithm types comprise a search boosting component subtype and a search pruning component; and
for each given satisficing planner component of the plurality of satisficing planner components, determining a component candidate, from among a plurality of component candidates, of the determined component type of the given satisficing planner component.
7. The computer system of claim 6 wherein the machine learning algorithm is a deep learning algorithm that includes a plurality of layers.
8. The computer system of claim 7 where the plurality of layers includes a plurality of hidden layers.
9. A computer program product comprising:
one or more computer-readable tangible storage media and program instructions stored on at least one of the one or more tangible storage media, the program instructions executable by a processor to cause the processor to perform a method comprising:
receiving a planner application data set that includes information indicative of contexts in which a newly-designed satisficing planner is expected to be applied;
receiving a machine learning algorithm that has been trained to design satisficing planners; and
applying the machine learning algorithm to the planner application data set to obtain a newly-designed satisficing planner design, with the newly-designed satisficing planner design including the information indicative of a plurality of satisficing planner components to be included in the newly-designed satisficing planner design and an order in which the satisficing components are to be run, wherein the applying of the machine learning algorithm includes the following operations:
for each given satisficing planner component of the plurality of satisficing planner components, determining a component type from a plurality of component types for the given satisficing planner component, wherein the plurality of component types are selected from a list consisting of heuristic function types, search algorithm types, and refinement algorithm types, and wherein the refinement algorithm types comprise a search boosting component subtype and a search pruning component; and
for each given satisficing planner component of the plurality of satisficing planner components, determining a component candidate, from among a plurality of component candidates, of the determined component type of the given satisficing planner component.
10. The computer program product of claim 9 wherein the machine learning algorithm is a deep learning algorithm that includes a plurality of layers.
11. The computer program product of claim 10 , wherein the plurality of layers includes a plurality of hidden layers.