A methodology for problem-formulation

Abstract

The process of operational research is traditionally described by a list of sub-processes, starting with a description, definition or formulation of the problem. Explicit methodology for this formulation is sparse. Earlier research on the nature of conflict and its resolution produced a diagrammatic notation, which seemed capable of extension and use as a problem-formulation 'language'. This thesis examines a wide range of methodology, directly or peripherally related to problem-formulation. It then argues the case for carrying out research to develop the diagrammatic notation as a medium for communication between an operational researcher and his clients, in order to establish in detail what their 'problem' is and perhaps to monitor the process of resolving it. The notation, and rules for its use, are developed and described in detail, and then used to help decide how the problem formulation process itself is to be managed. The difficulty of 'proving' a methodology is examined and the principles and purpose of experiment in this context is discussed. A suitable group of people with a shared problem was approached, and agreed to an experimental consultancy. The methodology was thereby tested 'successfully', in the sense defined in the thesis. The consultancy is in abeyance, but a start has been made with the use of the methodology in a study in support of a hospital accident and emergency department; it is also to be used in a study of decision-making in higher military commands. Further applied research will also be needed to find out how best it can be used to enhance the early stages of systems practice (Checkland), the analysis of options (Radford; Howard) and hypergame analysis (Bennett). Present evidence suggests that the procedures and practice described in this thesis are consistent with and complementary to these and other methodologies.