Why it is computationally harder to reconstruct the past than to predict the future (Q1376490)

This paper brings nothing new except a kind of physical interpretation of recent results concerning NP-hardness of computing exact bounds on solutions of linear interval equations. Considering the relationship between the vector of state variables \(x\) and the vector of ``future'' values \(y\) in the linear form \(y=Ax\), and assuming interval uncertainty of data, the authors obtain the results claimed in the title from the facts that given interval \(x\), computing \(y\) is polynomial-time whereas given interval \(y\), computing \(x\) is NP-hard.