Speaker: Amin Karimi – (Optima/Monash University)

Bio:

Amin Karimi is a Research Fellow in the Department of Data Science and Artificial Intelligence at Monash Information Technology. His research focuses on developing optimisation models and algorithms, reinforcement learning models, and machine learning algorithms across diverse fields such as construction, healthcare, energy, logistics, and supply chain management. Additionally, he has supervised several MSc data science projects at The University of Melbourne. Amin has received numerous awards, including the Best Student Paper Award at the 12th International Conference on City Logistics in France in 2023, as well as Best Paper Awards at the 6th International Symposium on Multimodal Transportation in China in 2024 and the 8th International Conference on Intelligent Transportation in Italy in 2024.

Abstract:

In early-stage building design, architectural, structural, and mechanical systems are often designed sequentially, with each discipline optimising its own decisions. However, individually satisfactory designs do not necessarily lead to a feasible or economical overall building. For example, structural beams and HVAC ductwork must share limited ceiling space, potentially creating conflicts that are discovered only after important design decisions have been finalised. Resolving these conflicts can require redesign, additional engineering solutions, increased costs, and project delays.

This research investigates how mathematical optimisation can support interdisciplinary coordination during early-stage building design. Through Building 4.0 CRC Project 71, we have developed optimisation models and a web-based decision-support platform that allows designers to generate, modify, and compare architectural, structural, and mechanical design alternatives. The framework integrates Constraint Programming (CP) for structural layout and thermal zoning with Mixed-Integer Programming (MIP) for cost-aware HVAC duct routing. It considers spatial constraints, airflow requirements, structural arrangements, ceiling clearance, and installation costs.

Computational experiments demonstrate how optimisation can move beyond individual disciplinary design to support interactive, coordinated decision-making, enabling designers to identify conflicts, understand trade-offs, and explore feasible alternatives before committing to a design.

This event is Online:
JOIN VIA ZOOM – MEETING ID: 873 1557 5255; PASSWORD: 778635

SEMINAR: WED 04 NOVEMBER 2026 4PM – 5PM (AEST, Melbourne Time)

Speaker: Dr Shahrzad Valizadeh – (Optima/RMIT University)

Bio:

Shahrzad Valizadeh completed her PhD in Business Information Systems at RMIT University, specialising in operations research, healthcare logistics, and machine learning-enabled optimisation. Her research focuses on data-driven decision support for perishable healthcare supply chains, particularly blood and platelet inventory management across hospital networks. She develops stochastic optimisation, reinforcement learning, approximate dynamic programming, and decision-aware machine learning models to improve forecasting, allocation, substitution, and transshipment decisions under uncertainty.

Abstract:

Healthcare supply chains are among the most operationally challenging optimisation environments, particularly when managing perishable medical products such as blood and platelets. Platelet inventory management is especially complex because demand is uncertain, units deteriorate quickly with age, and both shortages and wastage can have serious clinical and humanitarian consequences.

This project develops a decision-aware learning framework for blood inventory management that integrates demand forecasting with downstream optimisation decisions. The proposed framework trains predictive models with awareness of their impact on inventory allocation, shortages, wastage, substitution, and transshipment decisions in a multi-hospital platelet supply network. By combining stochastic optimisation, machine learning, and healthcare operations research, the project aims to deliver transparent and actionable decision support for clinical inventory managers.

This event is Online:
JOIN VIA ZOOM – MEETING ID: 873 1557 5255; PASSWORD: 778635

SEMINAR: WED 11 NOVEMBER 2026 12PM – 1PM (AEST, Melbourne Time)

Speaker: Dr Lele (Joyce) Zhang – (Optima/Melbourne University)

Bio:

Dr Lele (Joyce) Zhang is an Associate Professor in School of Mathematics and Statistics at the University of Melbourne (UoM). She has a Bachelors Degree in Engineering (Southeast University, China), and a PhD in Operations Research (UoM). Her areas of expertise are broadly modelling and optimisation for complex systems, traffic flow theory, Monte Carlo simulation, scheduling theory, City Logistics, and automated optimisation using large language models and machine learning. She was an Associated Investigator of ARC Centre of Excellence for Mathematical and Statistical Frontiers (ACEMS), and is a Chief Investigator of ARC Training Centre in Optimisation Technologies, Integrated Methodologies, and Applications (OPTIMA) and a co-founder of Physical Internet Lab at UoM. She also has extensive collaborations with industries in logistics, transport, and healthcare, including Australia Post, Department of Transport (Victoria), and Australian Red Cross Lifeblood. Her work has received several awards, including the second prize of Social Science Research Excellence Award by Lanzhou Province, China, the best poster and paper awards from International Physical Internet Conference (2025), and the best conference paper and presentation from International Symposium of Multimodal Transportation (2024).

Abstract:

Last-mile parcel delivery illustrates a difficult logistics challenge: decisions about when and where to deliver interact with vehicle routes, capacity, service commitments and customer preferences. The classical vehicle routing problem (VRP) helps delivery operators consolidate spatially by serving nearby customers together, but parcels for the same recipient may arrive at different times, creating opportunities for temporal consolidation. This talk develops that opportunity from deterministic planning to uncertain demand and customer choice. First, a two-echelon VRP formulation lets the operator jointly choose parcel schedules, home or locker delivery, and routes within delivery deadlines. The proposed bi-objective model captures the trade-off between operating cost and service satisfaction, while a three-phase heuristic tackles the integrated problem. Experiments on a Melbourne network show when holding parcels reduces repeat trips, particularly when travel is costly or customers are dispersed. Yet operators cannot simply change delivery times or locations, and parcel demand is uncertain when plans are made. A distributionally robust bilevel model therefore links operator incentives to customers’ utility-based choices; a tractable reformulation and three-level matheuristic connect those choices to routing and capacity. Computational results identify when deferral and mode switching complement each other, and how planning for uncertainty improves capacity reliability without excessive conservatism. These developments illustrate how optimisation can support more realistic logistics decisions.

This event is Hybrid:
JOIN VIA ZOOM – MEETING ID: 873 1557 5255; PASSWORD: 778635
Level 8, Room 8109, Melbourne Connect

SEMINAR: WED 21 OCTOBER 2026 4.00PM-5.00PM (AEST, Melbourne Time)

Speaker: Professor Sean Arisian – (Optima/Latrobe University)

Bio:

Sean Arisian is the founding leader of the Sustainable Operations discipline at La Trobe University. He is also an Associate Investigator with the ARC Training Centre in Optimisation Technologies (OPTIMA) and a Research Fellow at the La Trobe Centre for Global Security. His research focuses on resilience, digital sustainability, and critical supply chains and infrastructure systems. Sean is a Chief Investigator on an Australia–India Cyber and Critical Technology Partnership (AICCTP) project funded by the Department of Foreign Affairs and Trade, which explores the use of AI and blockchain technologies to strengthen the cyber resilience of critical energy infrastructure supply chains and support strategic technology governance.

Abstract:

This study presents a novel framework that combines cyber threat intelligence sharing with flexible cooperative procurement to help critical supply chains respond more effectively to cyber-attacks. Through a government-moderated Cyber Threat Intelligence Platform (CTIP), competing organizations can share intelligence, coordinate responses, and strengthen cybersecurity investments while maintaining commercial competitiveness. Drawing on a case study of Australia’s health and medical procurement sector, the findings show that flexible procurement arrangements can enhance operational resilience, improve economic welfare, and support continuity of supply during disruptions. The research highlights the important role of government in fostering collaboration and demonstrates how targeted interventions can strengthen cyber readiness across critical sectors such as healthcare, pharmaceuticals, and food supply chains.

This event is ONLINE ONLY:
JOIN VIA ZOOM – MEETING ID: 873 1557 5255; PASSWORD: 778635

SEMINAR: WED 07 OCTOBER 2026 4.00PM-5.00PM (AEST, Melbourne Time)

Speaker: Prof Tom Wilkening – (Optima/Melbourne University)

Bio:

Tom Wilkening is Professor of Economics at the University of Melbourne. His areas of research are in experimental economics, behavioural economics, and market design. In his research, Professor Wilkening explores how to design markets, matching systems, auctions, and economic institutions in environments where behavioural biases and inexperience may affect how people make decisions. His research combines experimental economics with theory to understand how markets evolve and to understand the role the government and private sector might play to improve them.

Professor Wilkening’s research has been published in the top general interest and field journals in his field including The American Economic Review, The Journal of Political Economy, The Journal of Economic Literature, The Journal of the European Economic Association, Management Science, The Journal of Public Economics, The Journal of Economic Theory, and Experimental Economics. His work has attracted significant support by the Australian research Council through a Discovery Early Career Research Award, a Discovery Grant, and a Future Fellowship Award. He is on the editorial boards of Experimental Economics and the Journal of the Economic Science Association.

Professor Wilkening is the director of MACTCH lab at the University of Melbourne, which combines market design with experimental economics and computer science to develop real-world matching systems and combinatorial exchanges. He leads a branch of the lab that investigates how combinatorial exchanges can be used to help farmers in developing countries consolidate fragmented landholdings.

Abstract:

Agriculture in low-income countries is characterised by misallocation of land across farmers and by fragmentation — the division of farms into smaller plots—which increases costs and limits farmers’ ability to adopt technologies with increasing returns. We explore whether a trading platform that enables conditional land swaps can help farmers address these challenges. Theoretically, combinatorial bidding can thicken markets, reduce exposure to risk, and facilitate coordination among traders. However, implementing such a platform presents two optimisation challenges. First, from an algorithmic perspective, determining both the set of winning bids and the prices at which plots should be traded requires solving constraint-based optimisation problems quickly enough to support near-continuous-time trading.

Second, from a user perspective, package bidding can make it difficult for farmers to identify mutually beneficial trades and formulate effective bids. We discuss our approaches to solving and scaling the underlying optimisation problems, as well as our ongoing work on bid suggestion mechanisms designed to help farmers identify and execute mutually beneficial trades.

This event is Hybrid:
JOIN VIA ZOOM – MEETING ID: 873 1557 5255; PASSWORD: 778635
Level 8, Room 8109, Melbourne Connect

SEMINAR: WED 28 OCTOBER 2026 4.00PM-5.00PM (AEST, Melbourne Time)

Speaker: Dr Mario Andrés Muñoz – (Optima/Melbourne University)

Bio:

Mario Andrés Muñoz Acosta is a Senior Research Fellow in the School of Computing and Information Systems at the University of Melbourne and a member of the ARC Training Centre in Optimisation Technologies, Integrated Methodologies, and Applications (OPTIMA). He received his PhD from the University of Melbourne in 2014, then held research fellow positions at Monash University and in the University of Melbourne’s School of Mathematics and Statistics before joining CIS and OPTIMA in 2022. His research studies the relationship between algorithm performance and problem structure, with a focus on landscape analysis, instance space analysis, and kernel methods for black-box optimisation. His current work develops distribution-based distances between optimisation problem instances.

Abstract:

Algorithm selection in black-box optimisation relies on describing the structure of the problem instance being solved. Exploratory Landscape Analysis (ELA) has provided the dominant approach: sampled evaluations are compressed into a fixed set of hand-crafted features, which are then used to relate landscape structure to algorithm performance. This feature-based view has been highly influential, but it also has well-known limitations. ELA features can depend on the benchmark suite, the transformation applied to the instance, and the sampling design, and they do not define a principled distance between problem instances.

This talk follows the development of alternatives to this feature-based pipeline. I will begin with early information-theoretic landscape features and their sensitivity to transformations and random evaluation grids. I will then discuss footprint-based analysis for comparing algorithm performance across instance space. The final part introduces LANDMARK, a representation that treats each landscape as a probability distribution and compares instances directly through a kernel-based discrepancy, without feature engineering. On the BBOB suite, this distance recovers meaningful instance-space structure and supports algorithm selection with accuracy comparable to ELA, while remaining interpretable and extensible to constrained and multi-objective settings.

This event is Hybrid:
JOIN VIA ZOOM – MEETING ID: 873 1557 5255; PASSWORD: 778635
Level 8, Room 8109, Melbourne Connect

SEMINAR: WED 30 SEPTEMBER 2026 4.00PM-5.00PM (AEST, Melbourne Time)

Speaker: Helena Ramalhinho Lourenço – (University of Barcelona)

Bio:

Helena Ramalhinho Lourenço is a Full Professor in the Economics and Business Department at Pompeu Fabra University (UPF) in Barcelona, Spain. She holds a Bachelor’s degree and a Master’s degree in Statistics and Operations Research from the Faculty of Sciences of the University of Lisbon, Portugal, and a PhD in Operations Research from Cornell University, New York, USA.

She has been involved in numerous research projects and consulting activities for business organisations in the fields of Operations Research and Logistics. Helena has published extensively in leading international scientific journals and has presented her research at international congresses and conferences.

Helena teaches across undergraduate, master’s and doctoral programs at UPF and other universities. She currently serves as Director of the Business Analytics Research Group at UPF.

Her research interests include Operations Research, Scheduling, Combinatorial Optimisation, Metaheuristics, Iterated Local Search, Heuristic Search Optimisation, Vehicle Routing, and applications in Supply Chain Management, Logistics, Operations Management, and Health Care.

Currently, Helena is the Vice-Rector (Vice-President) for Internationalisation at Pompeu Fabra University.

Abstract:

Optimisation plays a vital role in generating social impact by enabling the efficient allocation of resources, supporting better decision-making, and increasing the effectiveness of various initiatives. In sectors such as healthcare and social services, the use of optimisation techniques can lead to significant cost savings, improved utilisation of resources, higher satisfaction among personnel and service users, and-most importantly-ensures that limited resources are directed to those who need them most. Ultimately, this fosters a more equitable and sustainable society.

In this talk, we will introduce optimisation tools grounded in mathematical modelling, algorithmic methodologies, and metaheuristics, with a focus on their application to Social Care Services. We will showcase real-world implementations involving organisations such as Banc del Moviment, Suara, and the Barcelona City Council.

The presentation will highlight specific applications, including:

• Logistics and scheduling for in-home social and healthcare services
• Optimisation of assistive technology service delivery
• Mobility solutions for social care provision
• Optimisation of shared door-to-door transportation for people with disability

We will discuss the mathematical programming models and metaheuristic algorithms used in these cases, emphasising the unique challenges and considerations of applying these tools in social care compared to more traditional domains such as manufacturing and retail

This event is Online Only:
JOIN VIA ZOOM – MEETING ID: 873 1557 5255; PASSWORD: 778635

SEMINAR: WED 14 OCTOBER 2026 4:00PM – 5:00PM (AEST, Melbourne Time)

Speaker: Kelvin Liu – (Optima/Melbourne University)

Bio:

Kelvin is an OPTIMA Postdoctoral Research Fellow at the University of Melbourne. He completed his joint PhD with the University of Melbourne and KU Leuven in 2025, where his research focused on applying instance space analysis to mathematical model formulations of multi-dimensional cutting and packing problems. He recently completed a three-month research contract with GenV on the delivery of its Early School Wave. His current research is on applications of AI in optimisation.

Abstract:

Generation Victoria (GenV) is a longitudinal child and parent cohort study run from the Murdoch Children’s Research Institute, in partnership with other organisations including the University of Melbourne. Its Early School Wave will send assessment teams into around 1,800 Victorian primary schools from mid-2028, measuring children across domains including cardiovascular function, bone health, lung capacity, cognition and motor skills. These measurements will form the basis of future research into how early-life indicators relate to chronic disease and other health conditions, so the wave needs to be delivered efficiently and at scale without compromising data quality.

The talk outlines how optimisation techniques could support this delivery. Over a three-month project, GenV and I identified three problems central to delivering the wave: scheduling children through assessments within a school, deploying teams across the school network, and managing consumables and biosample logistics. I explored a range of scheduling and assignment methods for tackling the first two, which GenV identified as the more pressing. I also set out the limitations of those methods, which stem mainly from the information and data currently available.

This event is Online Only:
JOIN VIA ZOOM – MEETING ID: 873 1557 5255; PASSWORD: 778635

SEMINAR: WED 09 SEPTEMBER 2026 4:00PM – 5:00PM (AEST, Melbourne Time)

Speaker: Frits de Nijs – (Optima/Monash University)

Bio:

Frits is a Research Fellow in the Department of Data Science and Artificial Intelligence at the Faculty of Information Technology, working on the translation of combinatorial optimisation technology to industry. He obtained his PhD from Delft University of Technology on the application of sequential decision-making planners to solve demand-side management problems for a distribution network operator. Since joining Monash, he has developed solutions with various industry partners for optimising virtual power plant operation, carbon-intensity aware HVAC control, industrial facility location and configuration, and crew rostering problems.

Abstract:

For industry partners to adopt our optimisation solutions, having a correct and efficient optimisation model of their business problem is only part of the solution. Full-fledged system implementations should ideally allow users to: 1) interact with the optimisation model to explore the impact of assumptions, 2) use the system to find and compare a diversity of solutions, 3) understand why no solutions exist when the model proves infeasible (and how this may be repaired).

This talk presents how MiniZinc enables us to develop such interactive systems through two real-world case studies: a hydrogen and ammonia supply network optimisation problem, and a remote facility crew (re)rostering problem. In the talk, I will present our team’s solutions for both problems, and discuss commonalities in our methodology and the features of MiniZinc that underpin them: separation of data and constraints and modular model implementations, using model cross-checking for validity, applying diversity-of-solutions objectives, and annotating models to allow FindMUS to explain unsatisfiable models.

This event is Online Only:
JOIN VIA ZOOM – MEETING ID: 873 1557 5255; PASSWORD: 778635

SEMINAR: WED 26 AUGUST 2026 4:00PM – 5:00PM (AEST, Melbourne Time)

Speaker: Nicklas Klein – (Optima/Monash University)

Bio:

Nicklas is a Senior Lecturer in the Department of Econometrics and Business Statistics at Monash University. He holds a PhD in Operations Research from the University of Bern, and was a visiting researcher in the Supply Chain Management Department at Rutgers University and the Department of Technology, Operations and Statistics at NYU Stern during his doctoral studies. His research focuses on developing models and solution approaches for challenging optimisation problems in areas such as manufacturing, project management, and supply chain management, using mathematical programming, exact and heuristic algorithms, and machine learning techniques.

Abstract:

An increasing share of value-adding activities in many companies across various industries is nowadays organised as projects. A project is typically composed of precedence-related activities that require time and capacities of scarce resources, such as personnel, funds, or equipment, for execution. A key component of planning such projects is project scheduling, i.e., the allocation of resources to activities over time. We consider the resource-constrained project scheduling problem (RCPSP), which comprises determining the start times of all project activities such that the precedence relations and resource availabilities are respected and the project completion time (makespan) is minimised. We propose a novel mixed-integer linear programming (MILP) model based on a novel type of sequencing variable. Our computational analysis indicates that the proposed model outperforms existing state-of-the-art models on several sets of benchmark instances. Additionally, we discuss extensions of the model to related deterministic and stochastic project scheduling problems.

This event is Online Only:
JOIN VIA ZOOM – MEETING ID: 873 1557 5255; PASSWORD: 778635

SEMINAR: WED 19 AUGUST 2026 4:00PM – 5:00PM (AEST, Melbourne Time)