Iterative Search Strategy Research Articles

The use of acoustic versus optic range sensors in manufacturing systems

Range information is very important for automatic control of manufacturing systems. Part location, edge detection, and motion guidance are just a few examples of problems in which range information is required. Through simple time‐of‐flight measurements acoustic systems can conveniently and reliably provide the distance between the sensor and the workpiece. Scanning with the acoustic transducer furthermore allows one to reconstruct the nearfield 3‐D environment. Difficulties arise primarily from the need that the workpiece surface should be reasonably parallel to the surface of the acoustic transducer. Iterative search strategies allow one to overcome this problem. Alternatively, optical range‐finding systems typically use triangulation strategies to measure the sensor/workpiece distance. Their depth resolution is comparable to that of acoustic transducers. The lateral resolution is better in optical systems. Furthermore, the orientation dependence of the workpiece/sensor surface is less critical than it is for acoustic transducers. Acoustic systems are less expensive than their optical counterparts. They also can survive in the sometimes harsh manufacturing environment. The advantages and disadvantages of optical and acoustical range finding sensors are discussed in the context of applied manufacturing problems.

An algorithm for engineering design optimization

AbstractIn this paper a new concept for development of algorithms for optimal design of engineering systems is presented. The basic idea is to use upper and lower bounds on optimum cost to develop iterative search strategies. The main feature of the concept is that it does not rely on one‐dimensional search to compute a step size at any design iteration. Implication of the feature is that the algorithms based on this concept require evaluation of constraint functions only once at any design iteration. This is highly desirable for optimal design of engineering systems because evaluation of functions for such systems is very expensive due to their implicit dependence on design variables. An algorithm based on the new concept is derived in the paper. Several new step sizes are introduced and their relation to proper reduced optimal design problems are presented. A new step size based on the constant cost requirement at some design iterations is introduced. Numerical aspects for the algorithm are also presented. Based on the new algorithm, a general‐purpose computer code GRP2 is developed. The code is used to solve several problems to gain experience and insight for the algorithm. Numerical experience with examples is discussed. It is concluded that algorithms based on bounding optimum cost have substantial potential for applications in optimal design of engineering systems.

The development of a highly interactive searching technique for morphs—Minicomputer operated retrieval (partially heuristic) system

The basic structure of MORPHS was described at the Fifth Cranfield Conference when emphasis was placed on the automatic root finding/role applying procedures based on linguistic principles. This skeleton remains, but it has since been clothed with relatively complex, iterative search strategies. The major strategy employs the normal Boolean operators either in the form of strings without bracketing or in the form of marked steps. The latter enables back-tracking to be carried out. Recycling of search strategies is only limited by the number of tracks available for storing searches on the disc unit before over-writing takes place. Printing-out does not interfere with searching. A second strategy treats a string of keywords as a compound word; if it fails to find it the compound is then decomposed into its elements, which are then automatically treated as a string connected by AND logic; incidentally this technique also acts as a bracketing technique, but without the inconvenience of physical brackets. SDI procedures have been evolved as a by-product. Ultimately it is hoped that it may be possible to develop a degree of question answering ability: this could be achieved in several ways. The simplest (based on the modifying effect of certain interrogatory elements, e.g. WHO) has been evaluated and works. With the exception of the SDI system which employs 16k, all of the above systems were developed and originally operated in 12k of core with 16 bit words on a minicomputer. Nevertheless, despite this severe restriction on space it has been possible to implement searching strategies for dedicated on-line searching which are at least as good and in many ways better (e.g. automatic synonym substitution) than many employed in the large time-sharing systems currently available.