💻 computer science

Software Framework for Anthropocentric Dispatch Management and Decision Support in Industry 5.0-Oriented Educational Processes

This paper presents a software framework for anthropocentric dispatch management in Industry 5.0 educational settings that integrates chronotype-aware optimization, adaptive criterion weighting, and intelligent decision support to significantly improve schedule quality, student well-being, and academic outcomes while ensuring fairness across all participant groups.

Original authors: Oleh Sytnik

Published 2026-07-08

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Oleh Sytnik

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a university as a massive, complex orchestra. For years, the person in charge of writing the schedule (the "dispatcher") has been like a conductor who only cares about one thing: making sure every musician has a chair, a sheet of music, and a time to play, without anyone bumping into each other. They treat students and teachers like simple resources, similar to how a factory manager might treat machines.

The result? The schedule works perfectly on paper, but in reality, it's a disaster. It might force a "night owl" student to take their hardest math class at 8:00 AM when their brain is still asleep, or stack three difficult subjects back-to-back, leaving everyone exhausted and stressed.

This paper introduces a new software framework called ADMS (Anthropocentric Dispatch Management System) that changes the conductor's approach. Instead of just managing resources, it starts managing people. It treats the educational process as a "sociotechnical" system, meaning it cares about the human biology and psychology of everyone involved.

Here is how the new system works, broken down into simple concepts:

1. The "Human" Algorithm (The New Conductor)

The old way of making schedules used standard computer logic. The new system uses a special "evolutionary" algorithm (a type of AI that learns by trial and error) designed specifically for humans.

Chronotype Awareness: Just like people have different sleep schedules, the system knows if a student or teacher is a "morning lark" or a "night owl." It tries to schedule their hardest tasks when their brain is naturally most active.
Cognitive Compatibility: Imagine trying to eat a heavy steak right after a spicy meal; it doesn't sit well. Similarly, the system knows that putting a creative art class right after a heavy logic puzzle might confuse the brain. It arranges classes so they flow smoothly, like a well-planned meal.
Psychological Comfort: It avoids "back-to-back" stress. It ensures there are breaks and that the workload is spread out evenly throughout the week, rather than crushing everyone on Monday.

2. The "Smart Assistant" (The Translator)

In the past, computers would spit out a schedule and say, "Here it is, take it or leave it." If the schedule was weird, the human dispatcher had no idea why.
The new system includes an Intelligent Assistant. Think of this as a translator that explains the computer's thinking in plain English.

Why this class? It can tell the dispatcher: "We put this class at 10 AM because 80% of the students in this group are morning people."
What-If Scenarios: If a teacher says, "I can't teach on Tuesdays," the dispatcher can ask the computer: "What happens if I move this class?" The computer instantly simulates the change and says, "If you do that, stress levels will go down, but you'll have to move two other classes."
Human in the Loop: The computer never makes the final decision alone. It presents the top options with explanations, and a human dispatcher must approve them. This ensures the computer doesn't make a "technically perfect" but humanly terrible mistake.

3. The "Feedback Loop" (The Teacher's Pet)

Most computer systems are set once and never change. This system is like a student who learns from their mistakes.

At the end of the semester, students and teachers fill out a survey about how they felt.
The system uses this feedback to "re-calibrate" its priorities. If everyone said they were too stressed, the system learns to prioritize "low stress" over "perfect room usage" for the next semester. It gets smarter and more tailored to that specific university over time.

The Results: A Happier Orchestra

The researchers tested this system at a university in Ukraine. They compared the new "human-focused" schedule against the old "machine-focused" manual schedules. The results were significant:

Less Stress: Students and teachers reported feeling significantly less stressed (a drop of nearly 35% on stress scales).
Better Grades: Academic performance improved by about 13%.
More Attendance: Fewer people skipped class (attendance went up by 14%).
Fewer Conflicts: The number of scheduling errors (like two classes in the same room) dropped by over 80%.
Fairness: The system didn't just help the "morning people"; it improved the schedule for everyone, including the "night owls," without making anyone else worse off.

The Bottom Line

This paper argues that we need to move from Industry 4.0 (which is about efficiency and automation) to Industry 5.0 (which is about putting humans back at the center).

The software doesn't just build a schedule; it builds a better day for students and teachers. It proves that when you treat people like humans with biological rhythms and psychological needs, rather than just data points, the whole educational process runs smoother, happier, and more effectively.

Technical Summary: Software Framework for Anthropocentric Dispatch Management in Industry 5.0-Oriented Educational Processes

1. Problem Statement

Traditional higher education institution (HEI) class scheduling is predominantly treated as a technical resource-allocation problem, focusing on conflict-free assignment of instructors, groups, and rooms to time slots. This approach, rooted in Industry 4.0 logic, often ignores the chronobiological and psychological characteristics of participants (students and instructors). Consequently, formally valid timetables may be pedagogically counterproductive, leading to cognitive misalignment, increased stress, and reduced academic performance. Furthermore, existing systems lack mechanisms for:

Adaptive Weighting: Dynamically adjusting optimization criteria based on participant feedback.
Decision Support: Providing explainable recommendations, "what-if" scenario analysis, and risk-oriented warnings to human dispatchers.
Integration: Seamlessly connecting disparate institutional data sources (LMS, HR, surveys) into a unified optimization model.

This creates a gap between automated scheduling capabilities and the requirements of Industry 5.0, which prioritizes human-centricity, resilience, and explainability.

2. Methodology: The Anthropocentric Dispatch Management System (ADMS)

The paper proposes the Anthropocentric Dispatch Management System (ADMS), a modular monolithic software framework that integrates data collection, optimization, and decision support.

2.1 Architectural Framework

ADMS follows Clean Architecture principles, separating the domain, application, infrastructure, and API layers. Key subsystems include:

Data Collection and Storage (DCS): Ingests curriculum plans, chronotype profiles (via MEQ questionnaires), cognitive compatibility matrices, and participant constraints.
Optimization Core (OC): Executes the Anthropocentric Method of Dispatching (AMD), an evolutionary algorithm.
Satisfaction Regression Adaptation (SRA): A service that learns criterion weighting coefficients from post-semester feedback.
Intelligent Assistant (IA): Generates local/global explanations, performs "what-if" analysis, and computes risk scores.
Presentation Subsystem (PS): Provides a human-in-the-loop interface for dispatcher approval and audit logging.

2.2 The AMD Optimization Algorithm

The core optimization maximizes a four-component objective function $F(x, w)$ :
$F(x, w) = w_1 F_{tech}(x) + w_2 F_{circ}(x) + w_3 F_{psych}(x) + w_4 F_{cogn}(x)$

$F_{tech}$ (Technical Correctness): Ensures hard constraints (no double-booking, room capacity) are met.
$F_{circ}$ (Circadian Alignment): Aligns class slots with the chronotype activity peaks of students and instructors, modeled via Gaussian activity functions.
$F_{psych}$ (Psychological Comfort): Minimizes cognitive load variance and uncomfortable transitions between disciplines.
$F_{cogn}$ (Cognitive Compatibility): Optimizes the sequence of disciplines based on a directed compatibility matrix $S$ .

Key Algorithmic Operators:

Chronotype-Aware Initialization: Seeds the population with timetables biased toward chronobiologically plausible slots.
CPC (Chronotype-Preserving Crossover): A day-wise crossover operator that preserves contiguous daily blocks to maintain circadian quality, rather than using standard positional crossovers.
AWM (Anthropocentric Weighted Mutation): Concentrates mutations on slots with the lowest anthropocentric quality scores.
Repair Procedure: Systematically resolves hard constraint violations while attempting to restore chronobiological alignment.

2.3 Adaptive and Decision Support Mechanisms

SRA Service: Uses Ordinary Least Squares (OLS) or Ridge regression to map objective function components to participant satisfaction scores. It updates the weight vector $w$ iteratively, projecting it onto a constrained simplex to ensure no criterion is fully suppressed.
Intelligent Assistant: Provides a seven-dimensional decision vector for each candidate timetable, including risk scores (conflict, stress, cognitive, change) and natural language explanations. It supports "what-if" scenarios (e.g., instructor unavailability, room changes) to predict impacts before approval.

3. Key Contributions

Unified Anthropocentric Framework: The first deployable software system (ADMS) that unifies multi-criteria optimization with adaptive weight learning and intelligent decision support, moving beyond static, resource-only models.
Novel Evolutionary Operators: Introduction of CPC and AWM operators specifically designed to preserve and enhance human-centric metrics (chronotype and cognitive load) within the genetic algorithm search space.
Adaptive Weighting via SRA: A mechanism to dynamically learn institution-specific criterion weights from empirical feedback, addressing the rigidity of pre-configured weights in existing systems.
Explainable Human-in-the-Loop Workflow: A structured workflow where the dispatcher reviews top candidates with detailed risk assessments and explanations, ensuring human oversight and contestability of algorithmic decisions.
Industry 5.0 Alignment: The framework explicitly operationalizes Industry 5.0 principles (human-centricity, resilience, sustainability) in the context of educational process management.

4. Experimental Results

The framework was validated using a quasi-experimental design at a Ukrainian technical HEI and through ablation studies on generated datasets.

4.1 Algorithmic Performance (Ablation Study)

Quality Improvement: The full AMD configuration achieved a mean integral quality of 0.84, an 18.3% improvement over the Baseline GA (0.71).
Convergence Speed: AMD reached the quality threshold ( $F > 0.75$ ) in 38.4 minutes, a 34% reduction in time compared to the Baseline GA (58.2 minutes).
Stability: The standard deviation of final results ( $\sigma_{final}$ ) decreased from 0.043 (Baseline) to 0.029 (AMD), indicating more consistent outcomes.
Synergy: The combined effect of CPC and AWM exceeded the sum of their individual contributions, demonstrating synergistic interaction.

4.2 Human-Centric Outcomes (HEI Comparison)

Comparing a control group (traditional manual scheduling) with an experimental group (AMD with SRA-adapted weights):

Satisfaction: Student satisfaction increased by 43.7% (3.2 to 4.6/5); Instructor satisfaction increased by 41.4%.
Stress Reduction: Perceived stress (PSS-10) decreased by 34.8% (22.1 to 14.4), shifting the average from a moderate-high stress range to a low-stress range.
Academic Performance: Improved by 13.5% (3.7 to 4.2/5).
Attendance: Increased by 14.1%.
Conflict Reduction: Schedule conflicts dropped by 83.3% (from 12 to 2 per week).
Fairness: The system improved circadian alignment for evening chronotypes (the most disadvantaged group) by 45% (0.51 to 0.74) without reducing benefits for morning groups, confirming uniform distribution of gains.

4.3 System Usability

Dispatcher Acceptance: 76% of AI-generated recommendations were accepted; 24% required manual override, validating the value of human judgment.
Explainability: 100% success rate in "what-if" scenario analysis; explanations were rated 4.2/5 for usefulness.

5. Significance and Claims

The paper claims that ADMS addresses a critical methodological gap in educational scheduling by transitioning from a purely technical, Industry 4.0 paradigm to an Industry 5.0-oriented sociotechnical approach.

Shift in Paradigm: It demonstrates that treating scheduling as a complex human-oriented problem, rather than a simple constraint satisfaction problem, yields statistically significant improvements in well-being and academic outcomes.
Adaptability: The SRA service proves that systems can learn and adapt to specific institutional cultures and participant preferences over time, rather than relying on static, developer-defined parameters.
Trust and Transparency: By integrating explainability and human approval gates, the system fosters trust in automated decision-making, a prerequisite for the adoption of AI in sensitive educational contexts.
Scalability: The modular architecture and use of standard REST APIs suggest the framework is deployable across diverse HEIs and potentially adaptable to other sociotechnical domains (e.g., hospital shift scheduling).

The authors note limitations, including the single-institution validation and the reliance on self-reported feedback for weight adaptation, but assert that the framework provides a robust foundation for future multi-institutional validation and the development of more advanced, personalized educational planning tools.