2022
DOI: 10.1002/aaai.12053
|View full text |Cite
|
Sign up to set email alerts
|

A survey of knowledge‐based sequential decision‐making under uncertainty

Abstract: Reasoning with declarative knowledge (RDK) and sequential decision-making (SDM) are two key research areas in artificial intelligence. RDK methods reason with declarative domain knowledge, including commonsense knowledge, that is either provided a priori or acquired over time, while SDM methods (probabilistic planning [PP] and reinforcement learning [RL]) seek to compute action policies that maximize the expected cumulative utility over a time horizon; both classes of methods reason in the presence of uncertai… Show more

Help me understand this report
View preprint versions

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1

Citation Types

0
3
0

Year Published

2023
2023
2023
2023

Publication Types

Select...
3

Relationship

0
3

Authors

Journals

citations
Cited by 3 publications
(3 citation statements)
references
References 80 publications
(125 reference statements)
0
3
0
Order By: Relevance
“…The main task in decision‐making algorithms is that it must be fast enough to have real‐time. According to [33], decision‐making systems for self‐driving require four criteria, rapidity in the planned decision, coherency for avoiding unnecessary actions, and providentness which means that the module should foresee how the situation will evolve after some time/manoeuvres and include it in the decision‐ making, and finally, the predictability.…”
Section: Discussionmentioning
confidence: 99%
See 1 more Smart Citation
“…The main task in decision‐making algorithms is that it must be fast enough to have real‐time. According to [33], decision‐making systems for self‐driving require four criteria, rapidity in the planned decision, coherency for avoiding unnecessary actions, and providentness which means that the module should foresee how the situation will evolve after some time/manoeuvres and include it in the decision‐ making, and finally, the predictability.…”
Section: Discussionmentioning
confidence: 99%
“…Bayesian Networks (BNs) are typically part of Directed Acyclic Graphs [30]. In general, Decision Networks combine BNs with additional node types for actions and utilities [31], whereas Dynamic Networks 'DNs' provide support for probabilistic reasoning and decision-making in systems with uncertainty, and offer the capability to integrate multiple decision criteria, making them highly effective for path planning tasks [32,33]. According to [34], BNs can be structured into two levels, the situation assessment level to infer the current situation state based on the risk assessment, and the decisionmaking strategy level to deduce the manoeuvring decisions.…”
Section: Probabilistic Methodsmentioning
confidence: 99%
“…The latter is based on the dynamic optimization of the DN according to the load change and the operation constraints of the DN during different phases. In addition, sequential decision-making (SDM) is a time-sequential multi-stage optimization problem, where a controller can interact with system to obtain various sequential decisions (strategies) to maximize gains or minimize losses [6]. The SDM problem is more complex than a series of multiple independent decision problems since the controller considers the long-term effects of its decisions [7].…”
mentioning
confidence: 99%