ABSTRACT
Since the early seventies, there has been an increasing interest in the linear L1 approximation. According to Bloomfield and Steiger ([14], pp. 33, 34), it has been given many names, such as Discrete L1 approximation [1, 2, 8], L1 solution of overdetermined linear equations [1, 3, 4], Linear discrete L1 norm problem [6], L1 norm minimization [22], Least absolute deviations (LAD) [14], Minimum sum of absolute errors (MSAE) [17], L1–Approximation [18], Least absolute value (LAV) [23], L1 linear regression [24], Least absolute value regression [28] and others.
