Software Systems Lab

The Software Systems Lab aims to develop new methods, techniques, mechanisms and tools for building intelligent software systems. These intelligent software systems will enable new generations of applications, products and services in the ever-expanding digital eco-systems in which software, hardware and people interact.

The Lab focuses on developing AI-based, inter-connected and cloud-enabled software solutions, involving autonomous decision-making, decentralised optimisation and self-adaptation mechanisms used in intelligent multi-agent systems and their applications in cyber-physical-social eco-systems.

Our work includes:

  • Building and managing new digital platforms (such as smart exchanges, multi-cloud brokers, IoT/IoX/XaaS - everything as a service).
  • Enabling and optimising smart infrastructure (including smart cities, intelligent transport systems, smart energy grids and Industry 4.0).
  • Automating and optimising decision-making based on large-scale distributed, incomplete and imprecise data (involving machine learning, cognitive analytics and multi-agent learning).
  • Creating novel applications (e.g. cyber-resilient enterprise systems, community-based energy management, collective traffic optimisation, cognitive personal assistants, and digital twins in smart factories).

Research focus and capabilities

The Software Systems Lab research in intelligent, cloud-enabled and interconnected software intensive systems and their applications includes:

  • Artificial Intelligence systems - research into distributed AI, intelligent agents, collective intelligence and multi-agent systems.
    • The focus is on automated negotiation and autonomous decision-making, market-based and incentive-based distributed optimisation, autonomic self-organisation, machine learning and adaptation mechanisms, and their applications in building and managing open, large-scale, distributed systems and intelligent infrastructure
    • Applications include smart clouds, smart energy grids, smart cities and Industry 4.0.
  • Cloud computing systems - research into cloud performance management and cost-effective systems in the cloud.
    • The focus is on quality-assured adaptive clouds and multi-cloud brokers, big data management, adaptive workflow management, autonomous service-based systems in the cloud computing environment.
    • Applications include smart enterprises, data science, traffic management, IoT and smart cities.
  • Networked software systems - research into interconnected software systems.
    • The focus is on adaptive software architectures, context-aware systems, software-defined service networking, and data analytics for continuous software delivery and deployment.
    • Applications include cyber-physical-social systems in general, and cooperative ITS, IoT, IoP (internet of people) and smart cities in particular.

Case studies

The Smart Cloud Broker

The Smart Cloud Broker suite of software tools is the result of research begun in 2009 by Prof Ryszard Kowalczyk (Research Director), A/Prof Bao Quoc Vo (Research Project Manager), and Dr Mohan Chhetri (Software Architect/Researcher). It allows cloud infrastructure consumers to compare the different Infrastructure as a Service (IaaS) offerings from various cloud service providers, and purchase the cloud configuration from the most competitive provider with the most appropriate specification that best meets the user's technical and business requirements. Each component of the Smart Cloud Broker provides a unique functionality that can be used individually or in combination with other components. It includes:

  • Smart Cloud Bench Profiler - real-time comparative benchmarking based on data analytics and machine learning
  • Smart Cloud Purchaser – policy-based IaaSLA automation with intelligent agents
  • Smart Cloud Manager – automated consumer-centric QoS management
  • Smart Cloud Marketplace - AI-based market optimisation for open IaaS exchange 

Research and industry partners for the Smart Cloud Broker included the Smart Services CRC, AARNet and Zimbani Pty Ltd. A trial to benchmark different cloud infrastructures was conducted through the AARNet partnership with several universities. As a result, AARNet was able to develop a new strategy with regard to their cloud service orientation.

The Smart Cloud Broker (in particular, the Smart Cloud Purchaser and the Smart Cloud Marketplace) has also been considered very valuable to another industry partner, the Suncorp Group, in understanding relevant strategic innovation initiatives such as intelligent agents and digital marketplaces. Swinburne’s strategic partner in the new Digital Innovation Centre, Wipro Limited, has also expressed their interest in integrating the Smart Cloud Broker into their product and service offerings, especially the Artificial Intelligence Platform, Wipro HOLMES.

Market-oriented Optimisation of Smart Embedded Electricity Network

Swinburne’s Software Systems Laboratory is collaborating with Piechowski Energy Pty Ltd on the development of an intelligent decision support and automation system to optimise the cost-benefits and operational efficiency of smart embedded electricity networks (SEENs). SEENs include embedded networks enhanced with distributed energy generation, embedded storage, and demand response management options.

Intelligent energy management systems use intelligent optimisation algorithms and new computational market-oriented modelling. These features will allow the project to economically optimise the capital and operational costs of distributed energy resources in smart embedded networks in different buildings and precinct designs. This includes optimising dynamic pricing strategies for pricing-based demand response management, as well as optimising control to minimise the peak load and cost of energy supply to the embedded network participants while maximising the economic benefit to the embedded network owner.

Other benefits include optimised use and cost of the distributed energy resources in SEENs for infrastructure owners and network operators both in capacity creation (planning) and capacity utilisation (operations). An effective intelligent system to manage energy resources could also manage network loads more efficiently, reducing peak demand and improving network reliability and resilience – which in turn would reduce operational and capital costs.

The project is being led by Dr Mirek Piechowski, director of Piechowski Energy Pty Ltd, and Dr Ryszard Kowalczyk, director of the Swinburne’s Software Systems Lab and Head of Intelligent Agent Technology and Smart Energy Management Research.

Facilities

The Lab has an access to Swinburne Research Cloud – a dedicated private cloud service for Swinburne researchers and collaborators. The Research Cloud includes Swinburne Cloud Computing Test-bed (SwinCloud) that has recently received a major upgrade as part of strategic investment and commitment by the University. It serves as a hosting, experimentation and collaboration platform for various intelligent systems, artificial intelligence, cloud computing, machine learning and smart infrastructure projects with research and industry partners in Australia and internationally.

It also operates the Smart Cloud Broker portal, including an online Smart CloudBench test-bed developed jointly with AARNet. CloudBench is linked with several cloud providers located worldwide, including Amazon EC2, Rackspace, GoGrid, Azure, Google, HP, Nectar.

Contact

Professor Ryszard Kowalczyk
Director of the Software Systems Lab

T: +61 3 9214 5834
E: rkowalczyk@swin.edu.au