Behavioral Universe Network (BUN): A Behavioral Information-Based Framework for Complex Systems

Ailiya Borjigin; Cong He; Wei Zhou

REVIEW 4 major objections 4 minor 1 cited by

Behavioral Universe Network (BUN): A Behavioral Information-Based Framework for Complex Systems

T0 review · 4 major / 4 minor · reviewed 2026-08-16 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper proposes that every behavior in a digital ecosystem can be represented as a subject acting on an object through an operation, with all three governed by a shared Behavioral Information Base.

desk verdict A clearly written but thin conceptual framework that restates Lampson's access-control model and blackboard systems under new terminology, with a fixable formal gap in its central validation rule. read the letter →

arxiv 2504.15146 v1 pith:2OSML7KI submitted 2025-04-21 cs.AI

Wei Zhou , Ailiya Borjigin , Cong He This is my paper · ORCID

classification cs.AI

keywords BehavioralUniverseNetworkInformationBaseAgent-Interaction-Behaviorformalismmulti-agentsystemsdigitalgovernanceadaptivepolicyenforcementsemanticinteroperability

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper tries to establish that any interaction in a digital ecosystem, no matter how complex, can be represented as a subject acting on an object through an operation, with all three governed by a shared Behavioral Information Base (BIB). The claim is that this triple representation, together with a validity rule that checks policy constraints on subject, object, and operation, gives a unified foundation for analyzing, predicting, and coordinating behavior in multi-agent systems. A reader should care because it offers a single vocabulary for behavior across domains, replacing siloed agent-and-object models with one structure that can in principle handle security, business rules, and semantic context uniformly. The paper argues this yields more accurate behavior analysis, strong adaptability to change, and cross-domain interoperability.

What carries the argument

The key machinery is the AIB triple plus its validity rule. Behavior is written $a = S : f(O)$, and a behavior is permitted exactly when $(S \models P_1) \land (O \models P_2) \land (f(O) \models P_3)$ holds, with $P_1, P_2, P_3$ policy constraints on the subject, object, and operation. The Behavioral Information Base is the repository that makes the rule executable: it holds behavior histories, rule sets, semantic descriptors of objects, and predictive models. The BIB also acts as a blackboard for information-driven triggers, so one updated record can propagate a cascade of coordinated behaviors. The framework's power rests on treating the triple as the universal unit of interaction and on putting all governing knowledge in one shared, semantically rich store.

What would settle it

Build two independently implemented BUN systems that share the same BIB schema and rules but use different internal representations of an object's semantics, then ask both to adjudicate the same behavior request; if they disagree on whether $P_2$ holds (for example, whether a document tagged 'confidential' may be shared with a partner agent), the claim that the shared BIB yields uniform behavior validity across implementations is disproven.

Watch

Extended reading notes

Core claim

The paper's central claim is that behavior is first-class: every behavior $a = S : f(O)$, where $S$ is the subject, $O$ the object, and $f$ the operation, and it is valid only when the subject satisfies condition $P_1$, the object satisfies $P_2$, and the operation on the object satisfies $P_3$. These checks form a policy envelope, and the Behavioral Information Base stores the behavior histories, rules, semantics, and models needed to run the checks and to trigger further behaviors. The paper contends that when subjects and objects reference this shared base, agents become informed and autonomous: they choose actions within the envelope, learn from recorded outcomes, and coordinate through a blackboard-like trigger and propagation mechanism. If true, this gives a common formal core for digital governance, security, and multi-agent coordination that does not require rewriting agents when rules change.

Load-bearing premise

The framework assumes every meaningful constraint or context in an interaction can be captured by the three predicate checks $P_1$, $P_2$, and $P_3$, and that all agents interpret the shared Behavioral Information Base in the same way; the paper itself defers ontology development and BIB scalability to future work.

Editorial extensions

If this is right

Every interaction in a BUN system can be logged as a behavior trace with semantic context, making behavior analysis and prediction a matter of querying and mining the BIB.
Rules can be updated centrally in the BIB and all agents comply without code changes, giving system-wide adaptability.
Objects annotated with semantic descriptors can be handled by agents that have never seen them, enabling automation and cross-domain integration.
A behavior chain where one agent's action updates the BIB and triggers others gives an event-driven coordination mechanism analogous to a blackboard system.
Policy compliance becomes a runtime property: behaviors that fail any of $P_1$, $P_2$, or $P_3$ are invalid by construction.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Extension: if the BIB is treated as a shared ontology plus a policy engine, the AIB rule gives a natural correctness criterion for agent protocols: a protocol is sound if every reachable interaction satisfies $(S \models P_1) \land (O \models P_2) \land (f(O) \models P_3)$.
Extension: the blackboard trigger mechanism suggests a testable scalability prediction: coordination cost in BUN grows with the rate of BIB updates and subscription matches, not with the number of agents, which could be checked in a simulated smart-city scenario.
Extension: the framework's reliance on shared semantics implies that BUN's benefits should degrade gracefully as agents' ontologies diverge; a benchmark measuring agreement on $P_2$ for the same object across organizations would estimate real-world interoperability limits.
The paper leaves unstated whether the BIB is logically centralized or physically distributed; a concrete distributed BIB protocol with eventual consistency would be needed before the cross-enterprise claims can be deployed.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The punchline: this is a position paper, not a technical result. It introduces a vocabulary—Behavior = S:f(O), the AIB formalism, the BIB—but the underlying ideas are already in Lampson's protection model, blackboard architectures, and semantic IoT. There is no new equation, algorithm, dataset, simulation, or proof. What is new is packaging: a unified framing of behavior as a first-class citizen governed by a shared knowledge base. That framing is coherent as a vision statement and might be useful for teaching or grant writing, but it does not constitute a demonstrated framework.

What the paper does well: it is honest about its scope. Section 6 openly lists ontology and scalability as future work, and the citations to Lampson, blackboard systems, and semantic-web work are appropriate. The writing is clear, the examples (smart traffic, smart grid) are concrete, and the conceptual flow is easy to follow. As a survey of how existing components could be assembled into a behavior-centric architecture, it is not without value.

Where the soft spots are: the formal core has a real temporal gap. The validity rule in §2.3 writes (S |= P1) ∧ (O |= P2) ∧ (f(O) |= P3) ⇒ Behavior is valid, but §3.3 says the check happens "before or during" the behavior. If f(O) is the post-state output, it does not exist before execution. The paper could fix this by treating P3 as a predicate on the operation f itself (which §2.3's text actually suggests), or by specifying a predictive model for the output. As written, the notation is misleading and the enforcement claim is unsupported. This is the most concrete and fixable flaw.

The other soft spot is the reliance on reference [8], the authors' unpublished 2012 book, for foundational definitions. That source contains no machine-checked proofs or independent reproducibility, so the foundation is hard to audit. The paper also claims "precise analysis" and "deep coordination" without any formal semantics or experimental comparison; those claims are asserted through scenarios, not evidence. There is also no treatment of heterogeneous agents interpreting the same BIB semantics—the authors acknowledge this as future work, but it is central to the interoperability claims.

Who is this for? A reader who wants a bird's-eye view of behavior-centric coordination patterns, or a reviewer of a more developed paper that builds on these concepts. I would not cite it as a technical result.

Recommendation for peer review: if this lands in a technical journal or conference, I would desk reject it in its current form because the contribution is conceptual only and the formal gap is load-bearing. However, if the venue accepts short position papers or workshop-style vision pieces, it could be sent to a referee—but the referee should demand a precise formalization and at least one implemented example. As it stands, I would not spend referee time on it at a research venue.

Referee Report

4 major / 4 minor

Summary. The paper introduces the Behavioral Universe Network (BUN), a conceptual framework built on the Agent-Interaction-Behavior (AIB) formalism, in which every behavior is represented as a triple Behavior = S:f(O) with subject S, object O, and operation f, and is regulated by a shared Behavioral Information Base (BIB). The paper proposes that a behavior is valid only when (S |= P1) and (O |= P2) and (f(O) |= P3) hold, and argues that this rule, together with information-driven triggers and the integration of models, data, and rules in the BIB, enables accurate behavior analysis, adaptability, and cross-domain interoperability. The text is primarily conceptual: it defines the BUN components, describes how agents might query and update the BIB, gives illustrative examples from smart cities, smart grids, and cybersecurity, and closes with a list of open challenges. No formal semantics, proofs, simulations, or measurements are provided.

Significance. If the framework were made rigorous and validated, the core intuition—treating behavior as a first-class entity coordinated through a shared information base—could be a useful contribution to multi-agent systems and digital governance, especially by connecting policy enforcement with behavior logging. The paper is clearly written, and the blackboard-system analogy in Section 4.1 is apt. However, the submission does not establish any of its central claims: there is no formal semantics for |= and the predicates P1–P3, no enforcement mechanism that is logically consistent, no empirical evaluation, and no comparison with existing access-control, semantic-web, or multi-agent coordination frameworks. The foundational definitions are cited to the authors' own unpublished prior work [8], which is not publicly available. The claimed advantages in Section 5 therefore remain hypothetical.

major comments (4)

[§2.3 and §3.3] The validity rule (S |= P1) and (O |= P2) and (f(O) |= P3) =) Behavior is valid cannot serve as a pre-execution policy envelope as described in Section 3.3. Since f(O) is the output of operation f applied to O, it does not exist before the behavior is executed; checking it after execution would allow invalid behaviors to have occurred, while checking it before execution would require a predictive model of f(O) that is neither specified nor defined. If the authors intend f(O) to denote the behavior specification or request rather than the post-state, the notation is misleading, and the requirements on |=, P1, P2, and P3 are still left entirely unspecified.
[§2 and §3] The central definitions of the framework—Behavior = S:f(O), the validity of behavior, and the claim that a subject is the 'bearer of behavior'—are cited to the authors' own prior work [8], a 2012 book that is not available to the reader and contains no machine-checked proofs, code, or falsifiable predictions. No axioms, model theory, or formal semantics for BUN are given in this manuscript. As a result, the claimed theoretical foundation cannot be independently verified from the submitted text, and the framework is at present a collection of informal definitions and analogies rather than a formal theory.
[§5] The three claimed advantages—accurate behavior analysis and prediction, strong adaptability, and cross-domain interoperability—are supported only by hypothetical examples and informal reasoning. No experiments, simulations, case studies, or baseline comparisons are reported, and no metrics are defined against which the claimed benefits could be measured. The central claim that BUN 'enables precise analysis and coordination of behaviors' is therefore not established by the evidence presented in the manuscript.
[§6] The paper itself lists domain-specific ontology development and scalable BIB infrastructure as future work. These are load-bearing requirements for the framework's core promises of semantic interoperability and large-scale coordination: if the shared semantics of objects and behaviors are not specified and the BIB cannot scale, the cross-domain and coordination claims are not supported. The manuscript provides no formal or empirical argument that these challenges can be met within the proposed framework.

minor comments (4)

[Abstract and throughout] The text contains numerous typographical and formatting errors, such as 'treatssubjects', 'theAgent-Interaction-Behavior', 'theBehavioral', and 'theBehavioral Universe Network (BUN)[8]framework'; the paper would need a careful copyedit.
[References] Reference [8] is incomplete and is not publicly available; the authors should provide full bibliographic details or a freely accessible version, since the manuscript relies on it for foundational definitions.
[References] References [26] and [27] do not appear to be cited anywhere in the body of the text.
[§3.1] The phrase 'as per the digital asset model' refers to a model that is not defined or cited in the paper; either it should be defined or the phrase should be removed.

Circularity Check

1 steps flagged · score 6.0 of 10

The AIB validity rule defines the policy-compliance guarantee it then claims to enforce.

self definitional [Section 2.3 (validity rule) and Section 3.3 (BIB runtime enforcement)]
"⟨S |= P1⟩ ∧ ⟨O |= P2⟩ ∧ ⟨f (O) |= P3⟩ =⇒ Behavior is valid. ... Before or during a behavior, the system can check: ... are the contextual conditions P3 met? ... Only if all are true does the behavior proceed, so the behavior is inherently policy-compliant by design."

The formal rule defines 'Behavior is valid' as the conjunction of P1, P2, and P3 on the subject, object, and the output f(O). The later claim that a proceeding behavior is 'inherently policy-compliant by design' therefore restates the definition of 'valid' rather than deriving a substantive property. The circularity is compounded by a symbol shift: the formal P3 is a predicate on the post-state f(O), while the enforcement text checks P3 as a pre-execution 'contextual condition.' For a state-changing operation, f(O) does not exist before execution, so the pre-check cannot evaluate the formal P3; the guarantee that invalid behaviors are prevented is obtained by silently replacing P3 with a different, pre-checkable condition.

full rationale

The self-definitional loop is confined to the validity rule/enforcement pair. The paper's other architectural claims—BIB as a blackboard, semantic object enrichment, event-driven triggers, and integration of models, data, and rules—are not presented as predictions derived from a fitted parameter; they are design proposals with independent conceptual content, explicitly drawing on blackboard systems, BDI, and policy-based control. The self-citation to [8] for the AIB definitions and the 'bearer of behavior' statement is bibliographically weak—[8] is the authors' own book with no machine-checked or externally falsifiable result—but the definitions are restated in the present paper, so the citation itself is not the load-bearing step. The load-bearing step is the validity rule, which defines the guarantee it then claims to enforce. Because that guarantee ('policy-compliant by design') underpins the paper's claimed deep coordination and rule-based governance, partial circularity exists. However, the framework's remaining adaptive and interoperability claims are not reduced to their inputs, so a score of 6 rather than 8 or 10 is appropriate.

Assumptions & free parameters 0 free parameters · 4 assumptions · 3 invented entities

The central claims rest on a series of domain assumptions: that interactions can be captured by S:f(O), that three predicate checks suffice for validity, that a shared BIB is achievable and consistent across domains, and that agents will comply with it. These are stated rather than derived, and the foundational definitions are drawn from the authors' own prior work [8] without independent verification.

assumptions (4)

domain assumption Every behavior can be represented as S: f(O), where S is a subject, O an object, and f an operation.
Stated in Section 2 as the foundation of AIB; no argument is given that all complex interactions in digital ecosystems can be captured by a single triple.
ad hoc to paper A behavior is valid only if (S |= P1) and (O |= P2) and (f(O) |= P3).
Introduced in Section 2.3 as the policy envelope; the sufficiency and completeness of these three predicates is assumed, not proven.
domain assumption The Behavioral Information Base can serve as a consistent, authoritative shared substrate across heterogeneous domains.
Assumed in Sections 3 and 4; the paper lists ontology development as future work in Section 6, acknowledging this assumption is unresolved.
domain assumption Subjects are intelligent and adaptive agents that consult the BIB and comply with its rules.
Defined in Section 3.1; coordination benefits depend on agents actually following BIB guidance rather than acting selfishly or inconsistently.

invented entities (3)

Behavioral Information Base (BIB)
purpose: Central repository for behavior histories, rules, semantics, and models; the coordination substrate of BUN.
No implementation, benchmark, or falsifiable prediction is provided; it is a named conceptual component whose effectiveness is asserted.
Behavioral Universe Network (BUN)
purpose: The proposed framework instance of the AIB formalism.
Defined in the paper and in reference [8]; no deployed artifact or measurable outcome is provided.
Agent-Interaction-Behavior (AIB) formalism
purpose: Formal grounding for BUN.
Its rules are stated as axioms; the formalism is not encoded in a proof assistant or verified system.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Behavioral Universe Network (BUN): A Behavioral Information-Based Framework for Complex Systems." pith.science (2026). https://pith.science/paper/2OSML7KI

@misc{pith2026250415146,
  author       = {Pith},
  title        = {Pith review of: Behavioral Universe Network (BUN): A Behavioral Information-Based Framework for Complex Systems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2OSML7KI}},
  note         = {Machine review of arXiv:2504.15146}
}

read the original abstract

Modern digital ecosystems feature complex, dynamic interactions among autonomous entities across diverse domains. Traditional models often separate agents and objects, lacking a unified foundation to capture their interactive behaviors. This paper introduces the Behavioral Universe Network (BUN), a theoretical framework grounded in the Agent-Interaction-Behavior (AIB) formalism. BUN treats subjects (active agents), objects (resources), and behaviors (operations) as first-class entities, all governed by a shared Behavioral Information Base (BIB). We detail the AIB core concepts and demonstrate how BUN leverages information-driven triggers, semantic enrichment, and adaptive rules to coordinate multi-agent systems. We highlight key benefits: enhanced behavior analysis, strong adaptability, and cross-domain interoperability. We conclude by positioning BUN as a promising foundation for next-generation digital governance and intelligent applications.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

AI Agent Architecture for Decentralized Trading of Alternative Assets
cs.AI 2025-07 conditional novelty 5.0 of 10

GoldMine OS orchestrates four AI agents to tokenize physical gold on a permissioned blockchain, reporting sub-1.2 s issuance, tight spreads, and fault-triggered halts, but its scalability and safety claims rest on sim...

Reference graph

Works this paper leans on

28 extracted references · 28 canonical work pages · cited by 1 Pith paper

[8]

Zhou Wei, Qu Yanwen Introduction to Online Behavioral Science, [2012]

work page 2012
[1]

Sukhwal, V., Kankanhalli, A.: Agent-Based Modeling for Digital Governance: A Systematic Review.Government Information Quarterly, 39(2), 101–144 (2022)

work page 2022
[2]

Wiley, Chich- ester (2009)

Wooldridge, M.:An Introduction to MultiAgent Systems(2nd edn.). Wiley, Chich- ester (2009)

work page 2009
[3]

MIT Press, Cambridge (2018)

Sutton, R.S., Barto, A.G.:Reinforcement Learning: An Introduction(2nd edn.). MIT Press, Cambridge (2018)

work page 2018
[4]

Artificial Intelligence 117(2), 277–296 (2000)

Jennings, N.R.: On Agent-Based Software Engineering. Artificial Intelligence 117(2), 277–296 (2000)

work page 2000
[5]

ACM Trans

Abowd, G.D., Mynatt, E.D.:Charting Past, Present, and Future Research in Ubiq- uitous Computing. ACM Trans. Computer-Human Interaction7(1), 29–58 (2000)

work page 2000
[6]

Addison-Wesley, Boston (2003)

Bishop, M.:Computer Security: Art and Science. Addison-Wesley, Boston (2003)

work page 2003
[7]

Journal of the ACM42(4), 741–843 (1995)

Kifer, M., Lausen, G., Wu, J.:Logical Foundations of Object-Oriented and Frame- Based Languages. Journal of the ACM42(4), 741–843 (1995)

work page 1995

Show all 28 references

[9]

ACM Operating Systems Review8(1), 18–24 (1974)

Lampson, B.W.:Protection. ACM Operating Systems Review8(1), 18–24 (1974)

1974
[10]

AI Magazine7(2), 38–53 (1986)

Nii,H.P.: The Blackboard Model of Problem Solving and the Evolution of Blackboard Architectures. AI Magazine7(2), 38–53 (1986)

1986
[11]

ACM Com- puting Surveys 12(2), 213–253 (1980)

Erman, L.D., Hayes-Roth, F., Lesser, V.R., Reddy, D.R.:The Hearsay-II Speech- Understanding System: Integrating Knowledge to Resolve Uncertainty. ACM Com- puting Surveys 12(2), 213–253 (1980)

1980
[12]

In: Proc

Corkill, D.D.:Hierarchical Planning in a Distributed Environment. In: Proc. of the 6th Int. Joint Conf. on Artificial Intelligence (IJCAI), pp. 168–175 (1979)

1979
[13]

In: Proc

Freedman, R.:Interaction and Concurrency in Blackboard Architectures. In: Proc. of AAAI-82, pp. 34–39 (1982)

1982
[14]

In: Proc

Sommer, R., Paxson, V.:Outside the Closed World: On Using Machine Learning for Network Intrusion Detection. In: Proc. of IEEE S&P 2010, pp. 305–316 (2010)

2010
[15]

Smith, R.G.:The Contract Net Protocol: High-Level Communication and Control in a Distributed Problem Solver.IEEETrans.Computers 29(12),1104–1113(1980)

1980
[16]

Autonomous Agents and Multi-Agent Systems3(3), 285–312 (2000)

Wooldridge, M., Jennings, N.R., Kinny, D.: The Gaia Methodology for Agent- Oriented Analysis and Design. Autonomous Agents and Multi-Agent Systems3(3), 285–312 (2000)

2000
[17]

Au- tonomous Agents and Multi-Agent Systems11(3), 387–434 (2005)

Panait, L., Luke, S.:Cooperative Multi-Agent Learning: The State of the Art. Au- tonomous Agents and Multi-Agent Systems11(3), 387–434 (2005)

2005
[18]

In: Proc

Schroeder de Witt, C.A., Foerster, J.N., Farquhar, G., et al.:Multi-Agent Common Knowledge Reinforcement Learning. In: Proc. of NeurIPS 2019, pp. 14357–14368 (2019). Behavioral Universe Network Framework Based on AIB 17

2019
[19]

Advanced Engineering In- formatics 22(4), 486–501 (2008)

Bandini, S., Manzoni, S., Vizzari, G.:Crowd Modeling and Simulation: The Role of Multi-Agent Simulation in Design Support Systems. Advanced Engineering In- formatics 22(4), 486–501 (2008)

2008
[20]

In: Proc

Laskov, P., Šrndic, N.:Static Detection of Malicious JavaScript-Bearing PDF Doc- uments. In: Proc. of ACSAC 2011, pp. 373–382 (2011)

2011
[21]

In: Proc

Parunak, H.V.D., Baker, A., Clark, S.:Practical and Industrial Applications of Agent-Based Modeling. In: Proc. of AAMAS 2009, pp. 53–60 (2009)

2009
[22]

IEEE Access8, 45684–45695 (2020)

Park, K.J., Yang, H., Park, S., Choi, J.:Reinforcement Learning-Based Resource Management for 5G Network Slicing. IEEE Access8, 45684–45695 (2020)

2020
[23]

In: Proc

Gama, K., Touseau, L., Donsez, D.:Towards Interoperable Platforms for Urban Smart Applications. In: Proc. of the 13th Int. Conf. on Web Engineering (ICWE 2013), LNCS 7977, pp. 24–35. Springer (2013)

2013
[24]

ACM Trans

Kolozali, S., Bermudez-Edo, M., Puschmann, D., Ganzha, M., et al.:Semantic In- teroperability in the IoT: An Overview of Standards and Frameworks. ACM Trans. Internet Technology19(2), Article 22 (2019)

2019
[25]

Government Information Quar- terly 39(2), 101664 (2022)

Sukhwal, V., Kankanhalli, A.:Agent-Based Modeling in Digital Governance Re- search: A Review and Future Research Directions. Government Information Quar- terly 39(2), 101664 (2022)

2022
[26]

Hackathons for Good

Maillart, T., Gomez, L., Lombard, E., Nolte, A., Pisano, F.: Computational Diplomacy: How “Hackathons for Good” Feed a Participatory Future for Multi- lateralism in the Digital Age. Philosophical Transactions of the Royal Society A 382(20240103) (2024)

2024
[27]

Government Information Quarterly 39(3), 101740 (2022)

Saura, J.R., Ribeiro-Soriano, D., Palacios-Marqués, D.:Assessing Behavioral Data Science Privacy Issues in Government AI Deployment. Government Information Quarterly 39(3), 101740 (2022)

2022
[28]

Addison-Wesley, Reading (1982)

Denning, D.E.:Cryptography and Data Security. Addison-Wesley, Reading (1982)

1982

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.

[1] [8]

Zhou Wei, Qu Yanwen Introduction to Online Behavioral Science, [2012]

work page 2012

[2] [1]

Sukhwal, V., Kankanhalli, A.: Agent-Based Modeling for Digital Governance: A Systematic Review.Government Information Quarterly, 39(2), 101–144 (2022)

work page 2022

[3] [2]

Wiley, Chich- ester (2009)

Wooldridge, M.:An Introduction to MultiAgent Systems(2nd edn.). Wiley, Chich- ester (2009)

work page 2009

[4] [3]

MIT Press, Cambridge (2018)

Sutton, R.S., Barto, A.G.:Reinforcement Learning: An Introduction(2nd edn.). MIT Press, Cambridge (2018)

work page 2018

[5] [4]

Artificial Intelligence 117(2), 277–296 (2000)

Jennings, N.R.: On Agent-Based Software Engineering. Artificial Intelligence 117(2), 277–296 (2000)

work page 2000

[6] [5]

ACM Trans

Abowd, G.D., Mynatt, E.D.:Charting Past, Present, and Future Research in Ubiq- uitous Computing. ACM Trans. Computer-Human Interaction7(1), 29–58 (2000)

work page 2000

[7] [6]

Addison-Wesley, Boston (2003)

Bishop, M.:Computer Security: Art and Science. Addison-Wesley, Boston (2003)

work page 2003

[8] [7]

Journal of the ACM42(4), 741–843 (1995)

Kifer, M., Lausen, G., Wu, J.:Logical Foundations of Object-Oriented and Frame- Based Languages. Journal of the ACM42(4), 741–843 (1995)

work page 1995