An investigation into the use of knowledge representation in computer-integrated manufacturing

In a certain automobile factory, batch-painting of the body types in colours is controlled by an allocation system. This tries to balance production with orders, whilst making optimally-sized batches of colours. Sequences of cars entering painting cannot be optimised for easy selection of colour and batch size. `Over-production' is not allowed, in order to reduce buffer stocks of unsold vehicles. Paint quality is degraded by random effects. This thesis describes a toolkit which supports IKBS in an object-centred formalism. The intended domain of use for the toolkit is flexible manufacturing. A sizeable application program was developed, using the toolkit, to test the validity of the IKBS approach in solving the real manufacturing problem above, for which an existing conventional program was already being used. A detailed statistical analysis of the operating circumstances of the program was made to evaluate the likely need for the more flexible type of program for which the toolkit was intended. The IKBS program captures the many disparate and conflicting constraints in the scheduling knowledge and emulates the behaviour of the program installed in the factory. In the factory system, many possible, newly-discovered, heuristics would be awkward to represent and it would be impossible to make many new extensions. The representation scheme is capable of admitting changes to the knowledge, relying on the inherent encapsulating properties of object-centres programming to protect and isolate data. The object-centred scheme is supported by an enhancement of the `C' programming language and runs under BSD 4.2 UNIX. The structuring technique, using objects, provides a mechanism for separating control of expression of rule-based knowledge from the knowledge itself and allowing explicit `contexts', within which appropriate expression of knowledge can be done. Facilities are provided for acquisition of knowledge in a consistent manner.

A study of adaptive load balancing algorithms for distributed systems

With the advent of distributed computer systems with a largely transparent user interface, new questions have arisen regarding the management of such an environment by an operating system. One fertile area of research is that of load balancing, which attempts to improve system performance by redistributing the workload submitted to the system by the users. Early work in this field concentrated on static placement of computational objects to improve performance, given prior knowledge of process behaviour. More recently this has evolved into studying dynamic load balancing with process migration, thus allowing the system to adapt to varying loads. In this thesis, we describe a simulated system which facilitates experimentation with various load balancing algorithms. The system runs under UNIX and provides functions for user processes to communicate through software ports; processes reside on simulated homogeneous processors, connected by a user-specified topology, and a mechanism is included to allow migration of a process from one processor to another. We present the results of a study of adaptive load balancing algorithms, conducted using the aforementioned simulated system, under varying conditions; these results show the relative merits of different approaches to the load balancing problem, and we analyse the trade-offs between them. Following from this study, we present further novel modifications to suggested algorithms, and show their effects on system performance.