No description available.

No description available.

No description available.

MOODS (Multiple Objective Optimisation in Data and control path synthesis) is a synthesis system which provides the ability to automatically optimise a design from a behavioural to a structural VHDL description. This thesis details two sets of enhancements made to the original system to improve the overall quality of the final hardware implementations obtained, and expand the range of the accepted VHDL subset. Whereas the original MOODS considered each functional unit in the target module library to be a purely combinational logic block, the 'expanded modules' developed for this project provide a means of implementing sequential multi-cycle modules. These modules are defined as technology-independent templates, which are inline expanded into the internal design structure during synthesis. This enables inter-module optimisation to occur at the sub-module level, thus affording greater opportunities for unit sharing and module binding. The templates also facilitate the development of specialised interface modules. These enable the use of fixed timing I/O protocols for external interfacing, while maintaining maximum scheduling flexibility within the body of the behaviour. The second set of enhancements includes an improved implementation of behavioural VHDL as input to the system. This expands the previously limited subset to include such elements as signals, wait statements, concurrent processes, and functions and procedures. These are implemented according to the IEEE standard thereby preserving the computational effects of the VHDL simulation model. The final section of work involves the development and construction of an FPGA-based real-time audio-band spectrum analyser, synthesised within the MOODS environment. This design process provides valuable insights into the strengths and weaknesses of both MOODS and behavioural synthesis in general, serving as a firm foundation to guide future development of the system.

Yes / Matrix metalloproteinases (MMPs) are central to cancer development and metastasis. They are highly active in the tumor environment and absent or inactive in normal tissues; therefore they represent viable targets for cancer drug discovery. In this study we evaluated in silico docking to develop MMP-subtype-selective tumor-activated prodrugs. Proof of principle for this therapeutic approach was demonstrated in vitro against an aggressive human glioma model, with involvement of MMPs confirmed using pharmacological inhibition.

Conceptual design is a vital part of the design process during which designers first envision new ideas and then synthesize them into physical configurations that meet certain design specifications. In this research, a computational approach is developed to assist the designers perform this non-trivial task of navigating the design space for creating conceptual design configurations. The methodology is based on combining empirical reverse engineering techniques with a graph-grammar approach. Accordingly, design knowledge is systematically extracted from past designs, formulated as procedural grammar rules, and employed in building new design concepts. The implemented system provides a theoretical framework for automatically searching conceptual design spaces and produces novel alternative configurations to real design problems. The application of the approach to the design of various electromechanical devices shows the method's range of capabilities, and how it serves as a comparison to human conceptual design generation and as a tool to complement the skills of a designer.

No description available.

No description available.

Call number: LD2668 .T4 1985 S574 / Master of Science

Assembly-oriented design offers great potential for product rationalisation, increasing productivity and reducing lead time and cost. It results in simpler and more reliable products which are less expensive to assemble and manufacture. To facilitate assembly-oriented design, an assembly-oriented CAD environment is needed to incorporate Design for Assembly (DFA) evaluation from an early design stage. Assembly planning should also be integrated to support the DFA evaluation. This thesis reports the results of research towards supporting such an assembly-oriented CAD environment. A novel approach has been used to deploy an Expert Assembler to support proactive DFA evaluation and assembly sequence definition. This is particularly useful, as designers are rarely if ever assembly experts. Based on the fact that there are several areas needing expert support in this assembly-oriented CAD environment, but that different areas have very different requirements and different knowledge is involved, the Expert Assembler deployed contains several separated modules. Each module is an expert agent devised to tackle a problem area that uses a suitable problem solving strategy, knowledge representation and reasoning method. This brings a number of advantages that are detailed in the thesis. The thesis presents systematical ideas for support proactive DFA, with the focus on support for part count reduction and assembly sequence generation. This is realised by three elements of the expert agents: Part Count Advisor, Starting Part Advisor, and Next Part Advisor. Part count reduction is usually based on dialogue with the user. There is little computational support for this issue in any of the DFA methodologies and related literature. This research fills the gap: it brings computational support for part count reduction from the early design stage. The work has also made new progress in assembly sequence generation. The Starting Part Advisor and the Next Part Advisor cooperate with each other and with the user to provide suggestions dynamically and transparently regarding base part and the most suitable next part selection in assembly sequence definition. Case studies were used to test the effectiveness of the Advisors.