Abstract
Reconstructability analysis is defined as the process of investigating the possibilities of reconstructing properties of overall systems from the knowledge of the corresponding properties of collections of their various subsystems. In two recent papers, sets of procedures were presented which developed solutions of two problems associated with this process: that of generating a structure systems which could serve as appropriate collections of subsystems, and that of evaluating a structure system with respect to its adequacy for representing an overall system.
Use of these tools on problems arising from an actual systems investigation, however, involves resolution of a number of inductive ambiguities. In the final analysis, it is necessary to recognize elements of arbitrariness in this resolution. The places where this arbitrariness creeps in all involve decisions by an investigator who is using the tools. Ultimately, interactive software can be developed to aid the investigator in making these decisions. This paper gives a comprehensive description of reconstructability analysis from the perspective of identifying all the major points where these decisions must be made.
Notes
† This work was supported by the National Science Foundation under Grant No. ECS-8006590