Quantum information science holds immense promise for new and enhanced technologies in communications, computation, and precision measurement, as well as being closely linked to the fundamentals of quantum mechanics. One approach to realising these goals is to encode information onto single photons, the individual quanta of light. In this thesis, I .make use of a source of pairs of photons based on spontaneous four-wave mixing in microstructured fibre, which generates correlated signal and idler photons equally spaced in frequency above and below a bright pump laser powering the process. This source is well suited to demonstrating small-scale quantum information tasks involving a few quantum bits due to its brightness ' (the pump power required to reach a high generation rate is lower than comparable sources using bulk crystals) and its high coupling efficiency into standard optical fibre and photon detectors. Also, the use of a microstructure allows the dispersion properties of the fibre to be precisely engineered, adding control over the photons' wavelengths and spectral properties. Quantum interference is observed between signal photons in a Hong Ou Mandel dip with a visibility of 73%, which is thought to be limited mainly by impurity in the spectral state. Also, polarization entanglement between a signal-idler pair is generated with a fidelity of 87%. The source is then used in an interferometric setting to demonstrate enhanced sensitivity to optical phase changes and path-lengths compared to what can be achieved with classical light, and interference fringes are observed with down to a sixth of the wavelength of the classical pump laser. Using two sources of entangled pairs and a 'fusion' gate, three and four photon entangled GHZ states are produced with fidelities of 75% and 67% respectively. A rotated version of the four photon state is used to demonstrate three logic gates in the one-way model of quantum computing, the Hadamard H, the T phase rotation, and the controlled-NOT. Finally, the entangled state is extended to graph states with more complex structures, using path degrees of freedom to : add extra qubits to photons. A graph-state error correcting code is demonstrated.

No description available.

No description available.

No description available.

No description available.

No description available.

Context provides computing environments with the ability to adapt the information they provide by implicitly deriving the user's requirements from the context that surrounds the user at any situation. While considerable work has been carried out on the application of context, the meaning of context is often misinterpreted and poorly defined. Most of the attempts to use context have centred on the physical elements of the environment, users, or devices. However, only a little work has put emphasis on the context of the interaction itself. Accordingly, it is believed that it is necessary to incorporate context in the interaction design process as it becomes more important in interactive systems. This thesis proposes a method for designing interactive systems, where the dynamic changes in contexts and normative activities are captured and specified in the form of a formal language, called Context-based Activity Specified Language (CASL). The Context-based Activity Design (CoB AD) method is aimed at enabling a systematic design that assists or guides researchers and system designers in the field of interactive systems design to put emphasis on the context of the system use and its interaction. This method is built on three key components, namely the Context Ontology Model (COM), context reasoning mechanisms and specification of information fields. It also consists of the Context State Transition (CST) model that represents the transitions of activity states, where the ordering of activity states is often dependent on the context reasoning rules. One possible way to resolve such a conflict in context reasoning is to incorporate the concepts of norms and information fields, borrowed from Organisational Semiotics. This offers a novel method for designing interactive systems that is capable of modelling information fields as a set of norms. This generates different sets of outcomes depending on the different set of norms applied, which is useful for designing interactive systems. A series of conflict resolution strategies is proposed in this study, which mainly differ in the type and preference context that are used to resolve potential conflicts in the overlapping information fields . The CoB AD method is demonstrated by applying it into three different scenarios: (1) the activity design in a smart home, (2) the citizen lifecycle design and (3) the workaround design in a healthcare design work process. The method of treating conflicting norms in multiple information fields is presented in detail. Finally, the conclusions, contributions and future work identified from this study are expected to benefit researchers.

The main focus of the dissertation is on the method of integration of causal and resource reasoning that can expedite problem solving by graphbased planners.  We present a representation of the lifted planned graph constructed by the MADPlan planner, which is populated with exclusivity relations and binary constraints.  Moreover, the lifted nodes that appear within the graph maintain references to classes of resources which constitute the domain of the uninstantiated variables appearing in the lifted nodes.  The exclusivity relations, binary constraints, and resource references are used to guide the search towards the extraction of executable plans. We discuss the search method used in MADPlan that achieves the extraction of grounded plans directly from lifted planning graphs.  The main characteristic of the search is that it can switch between two different backtracking methods.  During chronological backtracking the algorithm tries a new resource allocation for a given node only when the node appears in the conflict set.  Otherwise the algorithm immediately backtracks further (backjumps) ignoring the remaining resource allocations for the last visited node.  The contributions discussed above allow our planner to scale very well with the increasing number of resources that appear in planning. The above contributions are direct extensions to the graphbased planning paradigm and they can be utilised by all graphbased planners.

The work defended in this thesis introduces a novel approach to constraint-centred qualitative reasoning in a non-constructive manner.  This new approach combines reasoning on the spectrum from fully qualitative to fully quantitative. In addition to this, all underlying algorithms have been implemented in parallel to decrease execution times. Previous work into parallel qualitative reasoning showed that execution time decreased over sequential implementations however the work had two main disadvantages.  First, not all stages of execution were implemented in parallel therefore the design was not optimal and secondly, the implementation was presented in the form of a dedicated hardware architecture. Several methods exist to reason with intervals or fuzzy numbers however no non-constructive approach offers results which are sound and complete.  A new qualitative reasoner, named JMorven, was implemented completely from scratch to overcome the limitations described above. JMorven is the successor to Morven but presents a novel set of algorithms working non-constructively and has an abstract parallel architecture which allows it to execute faster when run in distributed computing environments.  The novel work presented and tested in this thesis consists of: 1. A novel portable parallel architecture allowing speed-up of every stage of execution; 2.  The use of authority variable in a non-construction environment; 3. More accurate simulation using <i>n</i>-th order Taylor Series; 4.  Incorporating several non-constructive fuzzy interval simulation techniques; 5. A method of simulating fuzzy intervals non-constructively which is asymptotically sound and complete; 6. Offering the ability to carry out numerical simulations non-constructively from a qualitative model description; and 7. A single non-constructive simulation engine which is capable of reasoning on the spectrum from fully qualitative to fully quantitative.

No description available.