A5077292148- Smart Grid Energy Management
- Reinforcement Learning in Robotics
- Auction Theory and Applications
- Optimization and Search Problems
- Advanced Bandit Algorithms Research
- Scheduling and Optimization Algorithms
- Electric Vehicles and Infrastructure
- Microgrid Control and Optimization
- Game Theory and Applications
- Building Energy and Comfort Optimization
- Game Theory and Voting Systems
- Transportation and Mobility Innovations
- Vehicle Routing Optimization Methods
- Smart Grid Security and Resilience
- Energy Load and Power Forecasting
- Resource-Constrained Project Scheduling
- Energy Efficiency and Management
- Computational Geometry and Mesh Generation
- Integrated Energy Systems Optimization
- Bayesian Modeling and Causal Inference
- Complexity and Algorithms in Graphs
- Evolutionary Game Theory and Cooperation
- Energy and Environment Impacts
- Operations Management Techniques
- Advanced Control Systems Optimization
Monash University
2020-2025
Australian Regenerative Medicine Institute
2023
Delft University of Technology
2012-2018
This paper proposes a novel reinforcement learning (RL) architecture for the efficient scheduling and control of heating, ventilation air conditioning (HVAC) system in commercial building while harnessing its demand response (DR) potentials. With advances automated management systems, this can be achieved seamlessly by smart autonomous RL agent which takes best action, example, change HVAC temperature set point, necessary to electricity usage pattern signals, with minimal thermal comfort...
Abstract The emergence of cooperation in decentralized multi-agent systems is challenging; naive implementations learning algorithms typically fail to converge or equilibria without cooperation. Opponent modeling techniques, combined with reinforcement learning, have been successful promoting cooperation, but face challenges when other agents are plentiful anonymous. We envision environments which a sequence interactions different and heterogeneous agents. Inspired by models evolutionary...
Driven by the global decarbonization effort, rapid integration of renewable energy into conventional electricity grid presents new challenges and opportunities for battery storage system (BESS) participating in market. Energy arbitrage can be a significant source revenue BESS due to increasing price volatility spot market caused mismatch between generation demand. In addition, Frequency Control Ancillary Services (FCAS) markets established stabilize offer higher returns their capability...
In domains such as electric vehicle charging, smart distribution grids and autonomous warehouses, multiple agents share the same resources. When planning use of these resources, need to deal with uncertainty in domains. Although several models algorithms for constrained multiagent problems under have been proposed literature, it remains unclear when which algorithm can be applied. this survey we conceptualize establish a generic problem class based on Markov decision processes. We identify...
Renewable power sources such as wind and solar are inflexible in their energy production, which requires demand to rapidly follow supply order maintain balance. Promising controllable demands air-conditioners heat pumps use electric a temperature at setpoint. Such Thermostatically Controlled Loads (TCLs) have been shown be able curve using reactive control. In this paper we investigate the of planning under uncertainty pro-actively control an aggregation TCLs overcome temporary grid...
The rapid adoption of residential solar photovoltaics (PV) has resulted in regular overvoltage events, due to correlated reverse power flows. Currently, PV inverters prevent damage electronics by curtailing energy production response overvoltage. However, this disproportionately affects households at the far end feeder, leading an unfair allocation potential value produced. Globally optimizing for fair curtailment requires accurate feeder parameters, which are often unknown. This paper...
Multi-agent planning problems with constraints on global resource consumption occur in several domains. Existing algorithms for solving Markov Decision Processes can compute policies that meet a constraint expectation, but these provide no guarantees the probability violation will occur. We derive method to bound probabilities using Hoeffding's inequality. This is applied two existing approaches computing satisfying constraints: Constrained MDP framework and Column Generation approach. also...
Intelligent autonomous agents, designed to automate and simplify many aspects of our society, will increasingly be required also interact with other agents autonomously. Where interact, they are likely encounter resource constraints. For example, managing household appliances optimize electricity usage might need share the limited capacity distribution grid. This thesis describes research into new algorithms for optimizing behavior operating in constrained environments, when these have...
Resource constraints frequently complicate multi-agent planning problems. Existing algorithms for resource-constrained, problems rely on the assumption that are deterministic. However, resource themselves subject to uncertainty from external influences. Uncertainty about is especially challenging when agents must execute in an environment where communication unreliable, making on-line coordination difficult. In those cases, it a significant challenge find coordinated allocations at plan time...
Algorithms that involve both forecasting and optimization are at the core of solutions to many difficult real-world problems, such as in supply chains (inventory optimization), traffic, transition towards carbon-free energy generation battery/load/production scheduling sustainable systems. Typically, these scenarios we want solve an problem depends on unknown future values, which therefore need be forecast. As problems their own right, relatively few research has been done this area. This...
The Resource Constrained Project Scheduling Problem consists of finding start times for precedence-constrained activities which compete over renewable resources, with the goal to produce shortest schedule. method Justification is a very popular post-processing schedule optimization technique which, although it not clear exactly why, has been shown work well, even improving randomly generated schedules those produced by advanced heuristics. In this paper, we set out investigate why works so...
In multi-agent route planning, there is a set of autonomous vehicles (agents), each with their own start and destination locations. Agents want to reach destinations as quickly possible while avoiding conflicts other agents. We present single-agent planning algorithm that finds an optimal conflict-free plan given reservations from higher-priority also approach constructs schedule along fixed path. order obtain low-cost plans, the paths must provide agents sufficiently different alternatives...
Strategic recommendations (SR) refer to the problem where an intelligent agent observes sequential behaviors and activities of users decides when how interact with them optimize some long-term objectives, both for user business. These systems are in their infancy industry need practical solutions fundamental research challenges. At Adobe research, we have been implementing such various use-cases, including points interest recommendations, tutorial next step guidance multi-media editing...
Renewable power sources such as wind and solar are inflexible in their energy production, which requires demand to rapidly follow supply order maintain balance. Promising controllable demands air-conditioners heat pumps use electric a temperature at setpoint. Such Thermostatically Controlled Loads (TCLs) have been shown be able curve using reactive control. In this project we investigate the of planning under uncertainty pro-actively control an aggregation TCLs overcome temporary grid...