DOI: 10.3303/CET25120068 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Paper Received: 15 May 2025; Revised: 22 September 2025; Accepted: 12 September 2025 
Please cite this article as: Buics L., Eisinger B., 2025, Scenario-Based Optimization of Public Service Workflows Using Control Theory and P-
graph Methodology, Chemical Engineering Transactions, 120, 403-408  DOI:10.3303/CET25120068 
  

 CHEMICAL ENGINEERING TRANSACTIONS  
 

VOL. 120, 2025 

A publication of 

 

The Italian Association 
of Chemical Engineering 
Online at www.cetjournal.it 

Guest Editors: Bing Shen How, Viknesh Andiappan, Denny K.S. Ng, Hon Loong Lam, Petar S. Varbanov 

Copyright © 2025, AIDIC Servizi S.r.l. 

ISBN 979-12-81206-21-2; ISSN 2283-9216 

Scenario-Based Optimization of Public Service Workflows 

Using Control Theory and P-graph Methodology 

László Buicsa,*, Boglárka Eisingerb 

aDepartment of Leadership and Marketing, Vehicle Industry Research Center, Széchenyi István University, 1. Egyetem sq., 

Győr 9026, Hungary 
bVehicle Industry Research Center Széchenyi István University, 1. Egyetem sq., Győr 9026, Hungary 

buics.laszlo@ga.sze.hu 

Administrative workflows in public services, such as university enrolment are increasingly recognized as 

complex systems characterized by nonlinear dynamics, interdependencies, and variable operational 

constraints. These characteristics often lead to inefficiencies in resource allocation and compromise process 

sustainability. This study develops a novel control-theoretic optimization framework that integrates feedback, 

adaptive, and predictive control strategies with the P-graph methodology to address these challenges. The 

proposed approach models enrollment workflows using system dynamics and P-graph-based network 

optimization, enabling structured representation, real-time control, and scenario-based decision-making. To 

assess sustainability and efficiency, the model simulates varying scenarios of student demand, digital 

transformation levels, and administrative capacity. Key performance indicators (KPIs) such as resource 

utilization, delay reduction, and process flexibility are used to evaluate the comparative effectiveness of Model 

Predictive Control (MPC), discrete-event control, and feedback control strategies. Results show that optimized 

scenarios achieved a 50 % reduction in CO₂ emissions and over 2 h reduction in average processing time, 

significantly improving sustainability. This research uniquely applies control-theoretic methods to public 

administration and extends the P-graph methodology beyond industrial domains. The outcome is a scalable 

decision-support framework for policymakers aiming to enhance the resilience and sustainability of 

administrative service processes under dynamic conditions. 

1. Introduction 

Nowadays, the sustainability of administrative processes has become a strategic issue for public and private 

organisations, as they need to operate not only in an efficient and customer-oriented way, but also in an 

environmentally and socially responsible way (Balassa Eisinger and Buics, 2024). This means, among other 

things, reducing unnecessary resource and bureaucratic demands, reducing workloads to avoid burnout, and 

integrating environmental considerations into process design. A further challenge is that administrative 

processes are often subject to fluctuating workloads (e.g. overloaded during peak periods), consist of many 

discrete steps - between which a smooth transition must be ensured to avoid errors and congestion - and involve 

complex decision structures (covering continuous resource management and discrete events). To address 

these challenges, modern decision support methods are needed. Despite advances in process optimization and 

control, few studies have examined how nonlinear control theory can be integrated with combinatorial process 

synthesis (P-Graph) to manage sustainability in public service workflows. This study aims to address this gap 

by utilizing P-Graph methodology and control theory approaches together. Control theory tools provide a 

promising solution by dynamically responding to process state changes through closed-loop feedback. For 

example, automated manufacturing can reduce energy consumption and environmental pressures (Henao-

Hernández et al., 2019), and supply chains can increase dynamic planning capability and resilience (Ivanov et 

al., 2018). Advanced control techniques also contribute to the alignment of economic, environmental and social 

goals in production planning (May et al., 2023), while deep learning-based forecasting enables more flexible 

resource allocation in cloud services (Rossi et al., 2025). Advances in control theory have enabled efficient 

control of nonlinear, multivariate processes (Qin and Badgwell, 2003), and nowadays the incorporation of direct 

403



economic objective functions can also improve performance (Amrit et al., 2013). Discrete event control focuses 

on decisions that occur between two process steps, ensuring the correct sequence of steps and error-free 

operation. Non-blocking optimal control can be implemented in very complex systems (Wonham et al., 2018). 

A new approach to system stabilization has also been developed: by combining Lyapunov stability and max-

plus algebra, Konigsberg (2015) has developed a method to stabilize the state of discrete event systems over 

the long term (Konigsberg, 2015). To integrate continuous and discrete decision making, hybrid management 

approaches are used that simultaneously address resource allocation (continuous) and task scheduling 

(discrete) decisions. Moreover, addressing sustainability challenges requires new approaches in complex, multi-

dimensional systems, and model-predictive and hierarchical governance plays a key role in energy and 

emissions management (Daoutidis et al., 2016). Implementing sustainability management systems requires not 

only technical but also organisational changes. A case study shows that the integration of traditional and 

sustainability control systems helps embed sustainability (Beusch et al., 2022), while research shows that 

control systems are only effective when formal design is aligned with users' values and practices (Johnstone, 

2019). New technologies also help to promote sustainable operations: for example, in data centres, significant 

energy savings can be achieved by dynamically shifting workloads and using local energy sources, while smart 

methods (e.g. fuzzy logic, neural networks) are effective in nonlinear systems (Lee, 2023). The P-graph 

methodology was originally designed for network synthesis in chemical processes, but in recent years it has 

been shown to be effective for sustainability and administrative processes as well. Tick (2007) found that P-

graph-based workflow modeling is a preferred alternative to traditional modeling methods because its structural 

axioms guarantee the correctness of the models (Tick, 2007).A new technique allows generating and evaluating 

implicit decision rules from expert rankings using the P-graph, which is useful in complex, multi-criteria decision 

situations where decision makers' preferences are not clear (Low et al., 2020). The P-graph has also been 

proven in the design of sustainable industrial technologies: it has been used to optimize a biomass supply 

network to meet EU emission limits at minimal cost (Kodba et al., 2024). Migo-Sumagang et al. (2024) combined 

the P-graph with Monte Carlo simulation to investigate the robustness of portfolios of negative-emission 

technologies, finding the optimal balance between cost and risk; a recent review shows that the method is also 

effective in the design of negative-emission technologies and carbon management networks (Migo-Sumagang 

et al., 2022). In a university case study, P-graph analysis helped balance administrative and research capacities 

(Aviso et al., 2019), highlighting critical system resources and linkages. The aim of this study is to present a 

sustainability-oriented decision support framework, integrating P-graph methodology and management theory 

tools, which can be applied to optimise public administration processes and contribute to a more resilient, flexible 

and environmentally aware public service operation. The aim of the study is to develop a control-theoretic-based 

optimisation framework that can improve the efficiency, sustainability and resilience of administrative processes, 

in particular university enrolment. The research investigates how feedback, predictive and discrete event-driven 

management strategies can be integrated with the P-graph methodology and how this integration can be applied 

under different scenarios, such as different student enrolment, digitisation levels or capacity workloads. The 

novelty lies in the extension of the application of control theory tools to public administration processes and in 

the use of the P-graph methodology as a structured decision support tool in the public services sector, going 

beyond its industrial applications. By comparing the effectiveness of different control strategies with quantitative 

indicators such as resource utilisation, waiting time reduction, process flexibility, the optimal approach to 

promote responsive and sustainable public administration will be identified. 

2. Materials and Methods 

This study investigates the university enrollment process as a representative administrative service that is 

structured, resource-constrained, and exposed to demand fluctuations. The process consists of steps from the 

intake of applications to eligibility verification, document handling, decision-making, and final registration. The 

steps exhibit nonlinear behavior due to bottlenecks, task interdependencies, and varying available resources 

such that the process is an ideal candidate for control-based modeling and scenario analysis. The case study 

was replicated in manual-intensive and digitized versions to examine process effectiveness under different 

technology and organizational setups. Five process scenarios were developed to represent realistic variations 

in digital maturity and process control. The manual-intensive process without structured optimization was used 

as the baseline scenario. The other scenarios were the fully digital process with fixed structure and no control 

(Digital Static), the digital workflow with reactive control adjustments based on real-time data (Feedback 

Control), the predictive control strategy using historical demand trends (MPC), and the event-driven workflow 

transitions managed by system triggers (Discrete Event Control). Each scenario was analyzed under typical and 

peak student enrollment conditions to examine how the process performance (for example delay, resource 

utilization, throughput) responds to varying operational pressures and control strategies. The administrative 

404



operations were modelled as dynamic systems where the inputs (applications), the outputs (e.g. completed 

enrollments), and internal states (queue sizes, resource state) evolve over time. 

• Feedback Control: Controls task scheduling and staff in relation to queue length or system burden. 

• Model Predictive Control (MPC): Forecasts peak demand based on historical data and pre-allocates 

resources. 

• Discrete-Event Control: Uses process triggers (application load thresholds) to direct tasks or invoke fallback 

processes. 

The control structure enables scenario testing under various conditions of uncertainty, latency, and availability 

of resources. The P-Graph model was constructed using the P-Graph Studio software, encoding tasks as O-

type nodes (operating units) and resources as M-type nodes (material unit. Logical combinatorics were applied 

to generate optimal process structures using the maximal structure generation and solution structure generation  

algorithms (Tick, 2007). For control implementation, MPC was applied using linear prediction models with a 

rolling horizon strategy. Feedback control relied on real-time delay data to reallocate resources dynamically, 

while discrete-event control was simulated to evaluate system behavior under stochastic arrivals. P-graph was 

used to determine all potential workflow structures under input-output constraints, to optimize routing and 

resource allocation for minimum redundancy and delay, and to support control strategy development by 

simulating how structural configurations respond to control interventions. By combining P-graph with control-

theoretic methods, the research modeled both the structural and dynamic aspects of the enrollment workflow. 

The scenarios were compared on qualitative and quantitative parameters such as average processing time per 

request, system throughput and queue delay, resource utilization at peak and regular loads, process flexibility 

(the ability to adapt to changing demands), control effectiveness (the stability and responsiveness of the 

process), and sustainability factors like energy intensity and workload allocation. Each control strategy was 

compared on the basis of whether it could mitigate overload, improve resource utilization optimization, and raise 

process resilience and robustness. The P-graph was employed to represent structural adaptation in all cases 

and to quantify benefits gained through reconfiguration. 

3. Calculation 

To enable comparison across control strategies, the university admissions process was modelled as a discrete-

time dynamic system, with administrative tasks being executed in sequence with measurable resource 

consumption and time delays. Key assumptions include that the process handles a finite volume of applications 

(A), which are processed in discrete rounds (batches or single events); resources (R) (staff, systems, etc.) are 

pooled across tasks and they are constrained by their availability; tasks exhibit nonlinear delay characteristics 

under heavy load, such as queue time increases disproportionately with application volume; and control 

methods aim to minimize overall delay and balance the utilization of resources.  

The formula used for the calculations delay and queueing is: 

𝐷𝑗 = ∑(𝑇𝑖 +
𝑄𝑖(𝑡)

𝑅𝑖
∗ 𝑇𝑖

𝑛

𝑖=1

) (1) 

The formula used for the calculation average system delay is: 

𝐷𝑎𝑣 =
1

𝐴
∑ 𝐷𝑗

𝐴

𝑖=1

 (2) 

where Dj refers to Total process delay per application; Ti: Average processing time for task i; Qi(t): Queue length 

for task i at time t following a Poisson distribution to reflect real-world demand fluctuations; Ri: Number of 

resources assigned to task I; Dav: Average process delay per application; A: Number of applications to process. 

Each control strategy (s) is evaluated against the baseline (b) in terms of delay reduction: 

𝑃𝐼𝑠 =
𝐷𝑏

𝐷𝑠
 (3) 

where PI refers to Performance Index relative to baseline and S refers to the scenario index; PIs>1 indicates a 

performance improvement. To reflect the ability of a process to adapt under control strategies, a flexibility score 

is defined based on the number of feasible reconfigurations (Nf) are generated by the P-graph under each 

scenario: 

𝐹𝑠 =
𝑁𝑓

𝑁𝑓,𝑚𝑎𝑥
 (4) 

405



where F refers to Flexibility score of process configuration; Nf: Number of feasible reconfigurations; Nf,max is the 

total number of feasible process variants for the most flexible (hybrid) scenario. The formula used for the 

calculation resource utilization (Ui) for task i under scenario s: 

𝑈𝑖 =
𝑇𝑜𝑡𝑎𝑙 𝑎𝑐𝑡𝑖𝑣𝑒 𝑡𝑖𝑚𝑒

𝑇𝑜𝑡𝑎𝑙 𝑎𝑣𝑎𝑖𝑙𝑎𝑏𝑙𝑒 𝑡𝑖𝑚𝑒
=

𝐴 ∗ 𝑇𝑖

𝑇𝑚𝑎𝑥 ∗ 𝑅𝑖
 (5) 

The formulas used for the calculations are for the energy consumption per task: 

𝐸𝑖 = 𝑃𝑖 ∗ 𝑇𝑖 ∗ 𝑛𝑖 (6) 

𝐸𝑡𝑜𝑡𝑎𝑙 = ∑ 𝐸𝑖

𝑁

𝑖=1

 (7) 

where Ei: Energy consumption of task i (kWh); Pi: Power consumption of device or activity used in task i (kW); 

Tmax: Maximum processing time allowed for task i (h); ni: Number of instances of task i in the process; N: number 

of tasks in the scenario process. The formula used for the calculations are for CO2 emission is: 

𝐶𝑂2𝑡𝑜𝑡𝑎𝑙 = γ ∗ 𝐸𝑡𝑜𝑡𝑎𝑙 (8) 

where γ: Grid-specific emission factor (0.366 kg CO₂/kWh); Etotal: Total energy consumed by the process (kWh). 

These indicators provide a basis for multi-criteria evaluation of each control strategy’s impact on process 

sustainability and efficiency.  

In this control-theoretic framework, manipulated variables include staffing levels, resource assignments, and 

scheduling rules. Controlled variables are average processing time, queue length, and energy usage. 

4. Results 

In Table 1 we can see the five process scenarios simulated under both regular and peak-load conditions to 

evaluate performance in terms of average delay (D), resource utilization (U), and flexibility (F), Energy intensity 

and CO2 emission. 

Table 1: Comparative Performance of Control Strategies 

Scenario 
Avg. 

Delay (D) 

PI (vs. 

Baseline) 

Flexibility 

Score (F) 

Max 

Utilization (U) 

Energy Use 

(kWh/1000 cases) 

CO₂ Emissions 

(kg) 

Baseline (Manual) 4.5 h 1.00 0.30 0.75 182.00 66.61 

Digital Static 3.2 h 1.41 0.40 0.82 128.42 47.00 

Digital + Feedback 

Control 2.6 h 1.73 0.65 0.90 
97.27 35.60 

Digital + MPC 2.2 h 2.05 0.70 0.88 103.90 38.02 

Digital + Discrete-

Event 2.1 h 2.14 0.78 0.92 
89.58 32.78 

 

The scenario-based analysis showed that using control theory enhances administrative process performance 

in a number of ways. Compared to the manual baseline, digitally enabled scenarios in every scenario reduced 

processing delay, increased resource utilization, and improved structural flexibility. Of the control strategies 

employed, Digital + Discrete-Event Control performed best overall with the lowest average delay (2.1 h), highest 

flexibility score (0.78), and maximum utilization (92 %). For sustainability purposes, control scenarios that were 

control-driven exhibited large reductions in energy intensity and CO₂ emissions. The Digital+Feedback Control 

and MPC scenarios exhibited balanced performance improvements with environmental benefits, while the 

Discrete-Event Control scenario exhibited lowest emissions (32.78 kg CO₂ per 1,000 applications) and highest 

energy efficiency (0.09 kWh/application). These results validate the effectiveness of combining control-theoretic 

methods and P-graph-based structural optimization in enhancing administrative process operational and 

environmental performance. By using P-graph, all the scenarios were compared based on the number of 

alternative process structures. The hybrid and event-driven scenarios generated the maximum alternative 

structures because they were the most flexible. Flexibility score (F) was directly proportional to less delay and 

improved responsiveness, and therefore structural flexibility directly impacts control effectiveness. P-graph also 

found inefficiencies in the baseline case such as redundant routing and resource bottlenecks, and made it easy 

to generate optimized variants reducing queue accumulation under heavy load. Utilization analysis showed that 

the baseline process was suffering from underutilization during early stages as well as overload in bottleneck 

406



activities during peak periods. In comparison, control methods had more equal use of resources. From the 

results feedback control dynamically allocated resources based on backlog levels, MPC planned ahead based 

on expected load, smoothing out peak demand periods, while discrete-event control outperformed both of them 

by triggering structural adaptations and task switches based on real-time system status. These approaches 

ensured smoother queue dynamics and prevented overload at the peaks of enrollment. As we can see on Figure 

1, besides performance and control effectiveness the research evaluated each context with regard to process 

sustainability by means of two measures of performance. The energy intensity was quantified as the total energy 

consumed per processed application. Optimised digital control processes (especially MPC and discrete-event) 

evidenced lower energy per unit output due to more efficient task scheduling, avoidance of idle resources, and 

minimisation of repetition and waste caused by bottlenecks.  

 

Figure 1: CO2 Emission and energy use comparison across five simulated control scenarios 

According to the findings, control-based process design not only improves efficiency but also directly contributes 

to sustainability goals. By reducing energy use, load balancing, and task reduction, these strategies support 

leaner paradigms in public administration, especially when complemented with structural modeling using P-

graph. The results indicate that all of the control methods performed better than the baseline on delay, with the 

discrete-event control performing best overall in responsiveness and utilization balance. All control strategies 

demonstrate distinctive strengths. Feedback control was best with low data requirements and continuous 

monitoring but slower to adjust under nonlinear development. MPC was best with structured workflows with pre-

established patterns of demand. Discrete-event control, due to its increased complexity of implementation, 

showed most flexibility, stability, and scalability, especially when combined with P-graph structural optimization. 

The outcomes suggest that the fusion of discrete-event control logic with structural modeling using P-graph is a 

fruitful hybrid strategy to sustainable administrative process management. In practice, administrative workflow 

engineering with the introduction of control theory enables proactive adaptation to demand variation, efficient 

allocation of resources in hybrid digital/manual environments, and scenario-based assistance in decision-

making for process redesign, automation, and policy analysis. 

5. Conclusions 

This study proposed a control-theoretic framework supported by structural modeling of P-graph to enhance 

administrative process effectiveness and operational sustainability scenario-by-scenario. Using the university 

enrollment process as a case study, five control conditions were evaluated including manual workflows to 

digitally optimized with feedback, model predictive, and discrete-event control strategies. The findings show that 

the integration of control theory in designing administrative processes has a dramatic decrease in mean 

processing delays, improves resource utilization, and increases flexibility in processes. Among all the strategies 

experimented with, discrete-event control, particularly in combination with P-graph-based structure optimization, 

was the most balanced among responsiveness, efficiency, and flexibility. In addition, the inclusion of 

sustainability measures, such as energy intensity and CO₂ emissions demonstrated that control-maximized 

workflows matter to environmental targets. The digital discrete-event scenario exhibited the least energy usage 

and carbon footprint, suggesting the feasibility of such methodologies to achieving operational and ecological 

sustainability in public services. This research was based on collected data and simulated process behavior, 

with data based generalized assumptions on task durations, energy profiles, and system parameters. Regarding 

limitations real-world variability, such as the behavioral factors, existing policy constraints, and unpredictable 

demand fluctuations were not fully captured. Moreover, cost implications of the implementation of advanced 

control mechanisms were not addressed in depth. Future research should focus on the development of a more 

advanced hybrid control model with more integration of P-graph logic into actual administrative systems. Real-

time system data integration could allow adaptive control mechanisms that evolve over time as operational 

requirements change. Further research into multi-objective optimization could better optimize performance, cost, 

and sustainability goals. Lastly, extending the framework to other administrative functions, like financial dealings, 

compliance administration, or student assistance systems could extend and validate the model's versatility in 

broader public sector transformations. 

0.00

100.00

200.00
Baseline (Manual)

Digital Static

Digital + Feedback
Ctrl

Digital + MPC

Digital + Discrete-
Event

Energy Use (kWh/1,000 cases)

CO₂ Emissions (kg)

407



Nomenclature

A - number of applications to process

Dav - average process delay per application 

Dj - total process delay per application 

Ei - energy consumption of task i, kWh 

Etotal - total energy consumed by the process, kWh 

F - flexibility score of process configuration 

N - number of tasks in the scenario process 

Nf - number of feasible reconfigurations 

Nf,max - the total number of feasible process 

variants for the most flexible (hybrid) scenario 

ni - number of instances of task i in the process 

PI - Performance Index relative to baseline 

Pi - power consumption of device or activity used 

in task iii, kW 

PIs>1 - indicates a performance improvement 

Qi(t) - queue length for task i at time t 

Ri - number of resources assigned to task i 

Ti - average processing time for task i 

ti - time required for task i, h 

Ui - utilization of resource i 

γ - grid-specific emission factor, 0.366 kg 

CO₂/kWh

References 

Amrit R., Rawlings J.B., Biegler L.T., 2013, Optimizing process economics online using model predictive control. 

Computers and Chemical Engineering, 58, 334–343. 

Aviso K.B., Chiu A.S., Demeterio F.P. III, Lucas R.I.G., Tseng M.L., Tan R.R., 2019, Optimal human resource 

planning with P-graph for universities undergoing transition. Journal of Cleaner Production, 224, 811–822. 

Balassa Eisinger B., Buics L., 2024, The potential of the P-graph for optimizing public service processes. Clean 

Technologies and Environmental Policy, 1–13. 

Beusch P., Frisk J.E., Rosén M., Dilla W., 2022, Management control for sustainability: Towards integrated 

systems. Management Accounting Research, 54, 100777.  

Daoutidis P., Zachar M., Jogwar S.S., 2016, Sustainability and process control: A survey and perspective. 

Journal of Process Control, 44, 184–206.  

Henao-Hernández I., Solano-Charris E.L., Muñoz-Villamizar A., Santos J., Henríquez-Machado R., 2019, 

Control and monitoring for sustainable manufacturing in the Industry 4.0: A literature review. IFAC-

PapersOnLine, 52(10), 195–200.  

Ivanov D., Sethi S., Dolgui A., Sokolov B., 2018, A survey on control theory applications to operational systems, 

supply chain management, and Industry 4.0. Annual Reviews in Control, 46, 134–147. 

Johnstone L., 2019, Theorising and conceptualising the sustainability control system for effective sustainability 

management. Journal of Management Control, 30(1), 25–64. 

Kodba A., Pukšec T., Duić N., 2023, P-Graph approach for the economical optimisation of biomass supply 

network that meets requirements on greenhouse gas emissions savings – A case study of rural areas. 

Journal of Cleaner Production, 416, 137937. 

Konigsberg Z.R., 2015, The relation between the stabilization problem for discrete event systems modeled with 

timed Petri nets via Lyapunov methods and max-plus algebra. Journal of Applied Mathematics and Physics, 

3(7), 839–845. 

Lee M.F.R., 2023, A review on intelligent control theory and applications in process optimization and smart 

manufacturing. Processes, 11(11), 3171. 

Low C.X., Ng W.Y., Putra Z.A., Aviso K.B., Promentilla M.A.B., Tan R.R., 2020, Induction approach via P-Graph 

to rank clean technologies. Heliyon, 6(1), e03083.  

May M.C., Kiefer L., Frey A., Duffie N.A., Lanza G., 2023, Solving sustainable aggregate production planning 

with model predictive control. CIRP Annals, 72(1), 421–424.  

Mayer B., Killian M., Kozek M., 2017, Hierarchical model predictive control for sustainable building automation. 

Sustainability, 9(2), 264.  

Migo-Sumagang M.V., Tan R.R., Promentilla M.A.B., Aviso K.B., 2022, Applications of P-graph to carbon 

management: A mini-review. Chemical Engineering Transactions, 94, 31–36.  

Migo-Sumagang M.V., Aviso K.B., Tapia J.F.D., Tan R.R., 2024, P-graph and Monte Carlo simulation approach 

for sustainable and risk-managed CDR portfolios. Energy, 310, 133228.  

Qin S.J., Badgwell T.A., 2003, A survey of industrial model predictive control technology. Control Engineering 

Practice, 11(7), 733–764.  

Rossi A., Visentin A., Carraro D., Prestwich S., Brown K.N., 2025, Forecasting workload in cloud computing: 

Towards uncertainty-aware predictions and transfer learning. Cluster Computing, 28(4), 258.  

Tick J., 2007, P-graph-based workflow modelling. Acta Polytechnica Hungarica, 4(1), 75–88. 

Wonham W.M., Cai K., Rudie K., 2018, Supervisory control of discrete-event systems: A brief history. Annual 

Reviews in Control, 45, 250–256.  

408


	0211.pdf
	Scenario-Based Optimization of Public Service Workflows Using Control Theory and P-graph Methodology