Random walk is a powerful tool for searching a network, especially for a very large network such as the Internet. The cover time is an important measure of a random walk on a finite graph, and has been studied well. For the purpose of searching a network, it is quite natural to think that multiple crawlers might cover a network faster than a single crawler. Alon et al. (2011) showed that a multiple random walk by k crawlers covers a network k times faster than a single random walk in certain graphs such as complete graphs, random graphs, etc., while the speeding up ratio is limited only to log k times in other graphs such as cycles and paths. This paper investigates a multiplex random walk by k tokens, in which k tokens randomly walks on a graph independently according to an individual transition probability. For the cover time of a multiplex random walk, we present new inequalities similar to celebrated Matthews' inequalities for a single random walk, that provide upper and lower bounds of the cover time by its hitting times. We also show that the bounds are tight in certain graphs, namely complete graphs, bipartite complete graphs, and random graphs.