I. Introduction
Various "things" have recently become connected to the internet via an internetworking arrangement called the "Internet of Things" (IoT), and the vast amount of data thereby collected has brought many innovations to society. To advance IoT, all the various kinds of "things" should have the processing capability to solve versatile combinatorial optimization problems, that is, finding the optimal solution from among a huge combination of solutions in processes like identification, recognition, and route search [1]. Moreover, the number of combination patterns increases exponentially in accordance with the magnitude of a problem. For that reason, the optimal solution cannot be found in real time, which is indispensable for embedded applications, by conventional sequential computing using the round-robin method. Therefore, this study aims to incorporate a method for solving combinatorial optimization problems into "things."