A linear programming approach to conjoint measurment and trade-off analysis

There has been increasing attention placed on decomposition techniques such as conjoint measurement and trade-off analysis. Existing algorithms for both approaches use non-linear estimation techniques. In this paper, a linear integration rule is proposed along with a linear programming algorithm which is used to estimate part-utilities by minimizing the sum of the absolute deviations subject to monotonicity constraints. The algorithm was tested on simulated data and empirically. It was found that the algorithm performed well, especially when data are consistent with initial assumptions.