A System of Systems Methodology

The starting point was a classification of problem- contexts according to:

1. The nature of the systems(s) embedding the problem of concern
2. The relationship between relevant stakeholders

. Systems were seen to lie on a continuum ranging from simple to complex and, following Ackoff's terminology, problem-contexts labelled 'mechanical' if they contained relatively simple systems and 'systemic' if they housed complex systems. Decision-makers could be in a

• Unitary relationship to one another If they agreed upon a set of goals
• Pluralist If their objectives differed.

Problem-contexts, it followed, could also exhibit a 'unitary' or 'pluralist' character. Bringing the two dimensions of 'systems' and 'decision- makers' together, to form a four-celled matrix, yielded a classification of problem-contexts as mechanical-unitary, systemic-unitary, mechanical-pluralist and systemic-pluralist. Some brief justifi- cation was provided for the choice of the two dimensions forming the matrix.

	Unitary	Pluralist
Mechanical	Assumes that it is relatively easy to establish clear objectives for the system in which the problem resides, so the context must be unitary. The system is represented in a quantitative model which simulates its performance under different operational conditions something only possible if the system is simple and the context mechanical. Classical OR Systems analysis and systems engineering are most appropriate for solving problems.	Mechanical-pluralist contexts responded to the kind of systems design method proposed by Churchman, and by Mason and Mitroff in their 'strategic assumption surfacing and testing' methodology.
Systemic	Cybernetic approaches, such as advocated by Beer and in the socio-technical systems literature.	The soft systems thinking proposed by Ackoff and by Checkland could minister to problems set in systemic-pluralist problem-contexts. For example, Ackoff's 'interactive planning' exhibited, through the participative principle, a method to cope with pluralism and, through the proposed design for a 'responsive decision system', an attempt to come to terms with systemicity.

The above analysis is:

• Of practical importance because it can assist problem solvers to choose an appropriate methodology for the particular circumstances they faced.
• Its theoretical message is that OR and the other systems-based methodologies, far from being in competition with one another, could be used in an informed way, as a set of complementary tools, to tackle a much greater range of problem types.