We define an infinite dimensional modification of lower-semicomputability of density operators by Gács with an attempt to fix some problem in the paper. Our attempt is partly achieved by showing the existence of universal operator under some additional assumption. It is left as a future task to eliminate this assumption. We also see some properties and examples which stimulate further research. In particular, we show that universal operator has certain nontrivial form if it exists.
PreliminariesKolmogorov complexity is the notion of actual information content of finite string in computational point of view. This notion has been proposed by Kolmogorov -Solomonoff -Chaitin in 1960s and used in various areas as a basic tool to represent descriptive complexity. On the other hand, Since Shor's algorithm [1] has been discovered, the research on quantum information has made a great progress and produced various proposals on application to quantum information technology.Quantum Kolmogorov complexity is one of these branches appeared in early 2000s. Several different definitions are proposed so far [2][3][4], and some applications to quantum information are recently emerging [5][6]. However, it seems that there is very little progress in this area despite a decade has passed since these suggestions have been made, and a number of elementary facts are still not investigated.In particular, relationships between them are not clarified. In classical domain, there are several definitions of descriptive complexity and some of them are known as equivalent notions (Levin's coding theorem). This theorem, in some sense, guarantees that these notions are reliable.It naturally leads us to the following question: can we find any good relationship between these quantum complexities? In particular, if it turns out that some of them are equivalent, it would be helpful to make these notions more reliable and more applicable to other research subject such like quantum information theory.We particularly have interest on those by Berthiaume et al.[2] and Gács [3] since they are the quantum extension of plain Kolmogorov complexity and universal semimeasure, respectively. Levin's coding theorem claims that prefix Kolmogorov complexity and universal semimeasure are equivalent, so they are expected to be nearly equivalent.For Berthiaume's definition, there are several results about fundamental facts such like its invariance and relation between classical complexity [7][8][9]. As compared to this, there are not so much subsequent research of Gács' approach, so we mainly treat his definition. * Most part of this research was carried out without knowing about Tadaki's work [15]. Quite recently, in March 2014, Prof. Tadaki draw our attention to his work and we noticed that there are substantial overlaps between Tadaki's work and ours. As far as we know, his work seems to be the first one providing an extension of Gacs' work to the infinite-dimensional setting. In the present paper, we tried to reflect Tadaki's results as possible as we can. Interes...