Library · Optimization & control
What Finite State Machines & Logic Control got wrong, from 23 theses
Summary written by a model from the records below: The evaluated records show that finite state machines and logic control frameworks frequently suffer from severe state and transition explosion during formal verification, determinization, and behavioral modeling. They also encounter critical limitations when handling continuous physical dynamics, complex branching execution, and legacy hardware logic implementations. These records come from graduate theses, 2021 to 2026, each labeled with its degree where the repository states it, and each links to its thesis. They were extracted by language models reading the full text, so treat each as a lead to read, not a verdict.
Plus 9 problems the authors left open, listed at the end.
Failure modes, by theses reporting them
Select a mode to jump to its records.
State space explosion undermines explicit-state model checking and formal verification
Exhaustive model checking and reachability analysis of communication protocols, Petri nets, and timed specifications frequently exhausted available memory when exploring state graphs. Unhandled non-determinism, fine-grained time discretizations, and complex continuous dynamics caused state spaces to grow beyond solvable bounds.
Considered and rejected
Considered and rejected: Rejected encoding fine-grained robot movement times directly into the Spectra LTL/GR(1) specification because time discretization caused severe state-space explosion.
Automation of Multi-Robot Choreography Planning in Automotive Production · Leibniz Universität Hannover · PhD
Considered and rejected
Considered and rejected: Rejected allowing nodes to make pseudo-random choices via state seeds or exposing handler non-determinism to the model checker due to massive state graph explosion.
Techniques for Integrating Erasure Codes and Model Checkers with Distributed Systems · University of Washington · PhD
Considered and rejected
Considered and rejected: Rejected standard analytical/matrix-equation solutions (e.g. reachability graphs) for stochastic Petri nets due to intractable state-space explosion on large complex systems.
Modelling safety critical systems with ageing components, with application to underground railway risk and hazards · Nottingham · PhD
Tried and failed
incremental bounded model checking with state reuse applied to formal specification verification. Outcome: worse than baseline. Reason: a single unsatisfiable command prevented execution reduction through slicing or state reuse
Improving formal reasoning for software systems using machine learning and symbolic approaches · UT Austin · PhD
Transition complexity and maintenance bottlenecks cripple flat and handcrafted state machine architectures
Traditional and flat state machines scale poorly for autonomous systems and clinical workflows because state transitions and handcrafted rules grow quadratically or exponentially. Managing unorganized state flags or graphical transition diagrams creates fragile control logic that resists adaptation and becomes incomprehensible.
Considered and rejected
Considered and rejected: Finite State Machines (FSMs) for complex autonomous AAV behavior, rejected due to geometric/exponential growth in state transitions (n^2 or 2^n), fragility, poor runtime adaptability, and reduced debugging transparency
AERO: A Neo-Darwinian Approach to Time-Energy Optimal Behavioral Robotics · Harvard · Master's
Considered and rejected
Considered and rejected: Rejected traditional sense-think-act rule-based architectures / FSMs because state transitions and rule sets grow quadratically (O(n^2)) and fail to generalize to unmodeled behaviors.
Agent for Autonomous Driving based on Simulation Theories · University of Trento · PhD
Considered and rejected
Considered and rejected: Rejected Hierarchical Finite State Machines (HFSMs) for agent cognitive control due to exponential transition complexity (states squared) and hard-coded lack of adaptiveness, selecting Behavior Trees instead.
A Methodology for the Virtual Quantification of Operator Workloads in Support of Partially-Automated System Analysis · Georgia Tech · PhD
Considered and rejected
Considered and rejected: Rejected handcrafting full dialogue trees/rules due to high human effort and being error-prone for large state spaces.
Automating the Generation of Goal-Oriented Dialogue Managers for Healthcare · University of Trento · PhD
Considered and rejected
Considered and rejected: Rejected single-level flat FSMs for complex clinical workflows in favor of hierarchical FSMs with isolated State Branches to prevent transition complexity explosion.
Physical AI via Hierarchical Decision Processes · Johns Hopkins · PhD
Considered and rejected
Considered and rejected: Rejected unorganized flat Boolean state flags scattered across scripts/callbacks for workflow management in favor of explicit hFSM representations.
Physical AI via Hierarchical Decision Processes · Johns Hopkins · PhD
Automaton determinization and constraint compilation trigger combinatorial state explosion
Converting nondeterministic automata to deterministic representations for protocol tracking and pattern matching leads to exponential state space inflation when dealing with variable domains and large alphabets. Encoding transition functions into constraints or exhaustively checking homomorphism mappings similarly exhausts system memory.
Tried and failed
abstract finite tree automaton synthesis applied to inverse constructive solid geometry synthesis. Outcome: infeasible cost. Reason: state space explosion during refinement when handling repeating geometric patterns
Inductive and Deductive Synthesis for Database Applications · MIT · PhD
Tried and failed
brute-force enumeration of surjective mappings applied to finite state machine homomorphism equivalence. Outcome: infeasible cost. Reason: testing all permutations causes combinatorial explosion
TOWARD A ROBUST COMPUTATIONAL SOLUTION FOR FORMAL VERIFICATION AND VALIDATION IN MODEL-BASED SYSTEMS ENGINEERING · Georgia Tech · PhD
Tried and failed
constraint compilation of state machine transitions applied to large deterministic finite automata. Outcome: infeasible cost. Reason: memory exhaustion from encoding large transition functions into constraints
Correct Programs, Executed Correctly: Verifying Specifications And Executions · Penn · PhD
Hardware logic implementations and generic synthesis heuristics suffer from rigidity and inefficiency
Hardwired logic circuits and ladder logic diagrams lack operational flexibility and cannot support dynamic reconfiguration without halting physical processes. Automated logic synthesis and state assignment heuristics also underperformed specialized baselines in critical path delay, power reduction, and operating efficiency.
Tried and failed
logic synthesis mapping for quantum circuit compilation applied to reversible permutation and Toffoli circuit benchmarks. Outcome: worse than baseline. Reason: general logic decomposition lacks domain-specific direct gate optimizations present in specialized baselines
Optimizing Quantum Compilers: Efficient and Effective Algorithms · EPFL · PhD
Lost to a baseline
Vivado baseline achieved better run-time critical path delay than Reticle on the control-oriented finite state machine (fsm) benchmark across states 3, 5, 7, and 9
Programming Abstractions and Efficient Compilation Techniques for Modern FPGAs · University of Washington · PhD
Lost to a baseline
Current control logic in Grand'Maison PSP baseline resulted in lower efficiency (85.75%) compared to the proposed algorithm with start-up cost (86.30%) or without start-up cost (86.36%)
Complex conditional branching and recursive state evaluation introduce heavy execution overhead
Unadapted polyhedral transformations, nested conditional lookup trees, and recursive state machine folding inflate execution wall time and cause pipeline delays. Joint state-dependent logic and dynamic locking mechanisms similarly introduce redundant checking overhead that slows down simulation convergence compared to simpler static approaches.
Tried and failed
direct polyhedral compiler optimization for high-level synthesis applied to hardware acceleration of benchmark C loops. Outcome: worse than baseline. Reason: direct unadapted polyhedral transformations generate complex control logic inflating execution wall-time
Tools for efficient Deep Learning · Imperial · PhD
Tried and failed
Nested conditional branching for parameter lookup applied to dynamic laser power modulation arrays. Outcome: too slow. Reason: Evaluating thousands of sequential conditional statements introduced massive interpreter and execution overhead during fabrication
3D diffractive optics for linear interconnects and nonlinear processing · EPFL · PhD
Tried and failed
recursive state machine folding for automata verification applied to pattern matching on small inputs. Outcome: worse than baseline. Reason: per-step recursive proof overhead and match checking exceeded the cost of simple execution
Correct Programs, Executed Correctly: Verifying Specifications And Executions · Penn · PhD
Tried and failed
hierarchical joint state-dependent engagement logic applied to multi-component nonlinear mechanical damper simulation. Outcome: worse than baseline. Reason: overestimated engagement ratios and converged more slowly than independent component checks
Vibration Reduction and Energy Harvesting using Motion-Rectified Tuned Mass-Damper-Inerters in Semi-Submersible Offshore Wind Platforms · Virginia Tech · Master's
Lost to a baseline
PDL Bypass Queue lock introduces a dynamic priority calculation and redundant state overhead compared to Sodor's static bypassing logic.
LANGUAGE–BASED TECHNIQUES FOR BUILDING TIMING CHANNEL SECURE HARDWARE–SOFTWARE SYSTEMS · Cornell · PhD
Discrete state logic fails to accommodate continuous physical dynamics and evolving conversational context
Discrete state machines and fuzzy logic struggle when applied to continuous power dynamics or systems subject to drifting noise and unmodeled behavior. Rule-based dialogue state managers also fail across multi-turn interactions by ignoring history and triggering repetitive conversational loops.
Considered and rejected
Considered and rejected: Rejected pure FL (fuzzy logic) and fixed-gain PID controllers due to lack of dynamics modeling, poor robustness under drifting noise variables, and excessive manual gain tuning.
Adaptive Process Control for Achieving Consistent Mean Particles' States in Atmospheric Plasma Spray Process · Virginia Tech · Doctorate
Tried and failed
keyword-based regular expressions for dialogue state management applied to multi-turn conversational role-play systems. Reason: ignoring conversational history caused unnatural turn transitions and state inconsistencies
Development and validation of spoken dialog system-based oral communication tasks in an ESP context · Iowa State · PhD
Tried and failed
gendered conversational agents for product recommendation applied to e-commerce purchase intention. Outcome: worse than baseline. Reason: dialogue state management failures and conversational loop errors
Can (S)He Close the Deal? The influence of Purchase Intention Through Gender-Assigned Artificial Intelligence · Virginia Tech · Master's
Tried and failed
discrete topological switching for moving target defence applied to power grid state obfuscation. Reason: discrete state transitions produce separable clusters easily identified by unsupervised clustering algorithms
Operational moving target defences for improved power system cyber-physical security · Imperial · PhD
Left open by the authors
Problems the authors named and did not get to.
Left open
Extend the fault injection framework to support multi-fault and cascading failure injection without exponential state space explosion. Blocker: No concrete algorithmic approach or reduction technique is specified for mitigating combinatorial state explosion.
Efficient Fault Injection for Exposing and Reproducing Failures in Cloud Systems · Johns Hopkins · PhD
Left open
Model high-order and long-term sequential dependencies in Markov chains for recommendation while mitigating state space explosion. Blocker: No specific approach, mathematical formulation, or mitigation technique is defined to handle the state explosion
Semantic Embedded Sequential Recommendation for E-Commerce Products through Mining Customers’ Historical Interactions and Products’ Data · University of Windsor · PhD
Left open
Develop non-exponential synthesis algorithms for Event-based STL specifications that avoid Büchi automaton state explosion. Blocker: No concrete algorithmic approach or mechanism is specified to overcome the exponential complexity.
Event-Based Signal Temporal Logic Synthesis for Single and Multi-Robot Tasks · Cornell · PhD
Left open
Develop scalable Markov model approximations for multi-body robotic swarms to mitigate exponential state-space growth. Blocker: None
Design, modeling, and optimization of robots exploiting fluid structure interactions · EPFL · PhD
Left open
Extend the SMT- and BDD-based fault-tolerant design space exploration framework to handle dynamic state evolution and dynamic faults. Blocker: The task is described as an abstract research direction without a concrete mathematical approach or target specification
Automatic Design Space Exploration of Fault-tolerant Embedded Systems Architectures · University of Trento · PhD
Left open
Develop long-horizon memory maintenance and consistent state representation mechanisms for robot manipulation perception and planning. Blocker: The task is an open-ended research direction without specific architectural requirements, metrics, or benchmarks defined.
Generalizable Robot Manipulation through Unified Perception, Policy Learning, and Planning · MIT · PhD
Left open
Develop a hybrid model-and-data-based fault identification technique to identify root-cause variables in extended dynamical state-space models. Blocker: Vague task specification without concrete formulations, targets, or algorithms
Fault Detection and Identification of Large-scale Dynamical Systems · MIT · PhD
Left open
Automate probabilistic separation logic verification by synthesizing probabilistic loop invariants for general programs. Blocker: Lacks specific algorithm design, target verification framework, or formal heuristics for invariant synthesis
PROBABILISTIC SEPARATION LOGICS FOR RANDOMIZED ALGORITHMS · Cornell · PhD
Left open
Evaluate encoded state check error detection on mobile autonomous systems in distributed dynamic traffic settings via simulation. Blocker: Lack of specific evaluation metrics, experimental setups, or target autonomous system models
REAL-TIME ERROR DETECTION AND CORRECTION FOR ROBUST OPERATION OF AUTONOMOUS SYSTEMS USING ENCODED STATE CHECKS · Georgia Tech · PhD
Who has worked on this
Authors of the public theses cited on this page, most-cited first.
- Yihao Liu · Johns Hopkins · PhD · 2026
Physical AI via Hierarchical Decision Processes - Eleftherios Ioannidis · Penn · PhD · 2025
Correct Programs, Executed Correctly: Verifying Specifications And Executions - Eric Roslin Wete Poaka · Leibniz Universität Hannover · PhD · 2026
Automation of Multi-Robot Choreography Planning in Automotive Production - Xiaolin Fang · MIT · PhD · 2025
Generalizable Robot Manipulation through Unified Perception, Policy Learning, and Planning - Jialu Bao · Cornell · PhD · 2025
PROBABILISTIC SEPARATION LOGICS FOR RANDOMIZED ALGORITHMS - Haoze Wu · Johns Hopkins · PhD · 2025
Efficient Fault Injection for Exposing and Reproducing Failures in Cloud Systems - Nana Obayashi · EPFL · PhD · 2025
Design, modeling, and optimization of robots exploiting fluid structure interactions - Francesco Gerini · EPFL · PhD · 2024
Advanced control strategies to exploit the hydropower potential enhancing ancillary services provision to the power system - Niyazi Ulas Dinç · EPFL · PhD · 2024
3D diffractive optics for linear interconnects and nonlinear processing - Lauren Elizabeth Hall · Virginia Tech · Master's · 2024
Vibration Reduction and Energy Harvesting using Motion-Rectified Tuned Mass-Damper-Inerters in Semi-Submersible Offshore Wind Platforms - Mahreen Nasir Butt · University of Windsor · PhD · 2024
Semantic Embedded Sequential Recommendation for E-Commerce Products through Mining Customers’ Historical Interactions and Products’ Data - Drew Zagieboylo · Cornell · PhD · 2023
LANGUAGE–BASED TECHNIQUES FOR BUILDING TIMING CHANNEL SECURE HARDWARE–SOFTWARE SYSTEMS
Checking a claim in this area?
We can run the same search on any method or claim. If nothing turns up, we will say so, and that proves nothing on its own.