Search

Abstract.

Contains reprints of articles published by members of the department.

This paper deals with a dynamic linear programming model on a pilot scale that attempts to describe in physical terms many of the technological interactions within and across the sections of the American Economy, including a detailed energy sector. The general aim of the model is to provide information on what the country would achieve over the long run in physical terms in the face of a changing energy picture. The model is expected to (1) incorporate most recent available data of sufficiently good quality (albeit in an aggregated form) that it can be used to generate some meaningful scenarios showing how the economy might be affected if the energy picture evolves in a specific way, and (2) provide us at the Systems Optimization Laboratory a prototype for research in solving large-scale linear programming models of energy systems.

One of the ultimate goals of robotics research is to create autonomous robots. Progress toward this goal requires advances in many domains, including automatic motion planning. The basic problem in motion planning is to construct a collision-free path for a moving object among fixed obstacles. Several approaches have been proposed, including cell decomposition, retraction, and potential field. Nevertheless, most existing planners still lack efficiency, or reliability, or both. In this paper, we consider one of the most popular approaches to path planning: hierarchical approximate cell decomposition. We propose a set of new algorithms for constructing more efficient and reliable path planners based on this general approach. These algorithms concern the hierarchical decomposition of the robot's configuration space into rectangloid cells, and the search of the connectivity graphs built at each level of decomposition. We have implemented these algorithms in a path planner and experimented with this planner on various examples. Some are described in the paper. These experiments show that our planner is significantly faster than previous planners based on the same general approach. (KR).