Active occurrence-matrix-based approach to design decomposition

An efficient algorithm is presented for decomposing constraint equations defining an engineering design. It is assumed that the design problem is modelled mathematically by m independent, nonconflicting algebraic equations (constraints) in n design variables. The design procedure is initiated by specifying (or inputting) ( n− m) variables, one by one. The algorithm uses an active occurrence-matrix approach to determine which variables can be computed, and the most efficient sequence for computing these variables. The procedure is independent of the order in which the input variable(s) are specified.