Presentation Information
[C-10_C-11-04]Bidirectional Pseudopod-state Control in an Analog Electronic Amoeba for Solving Boolean Satisfiability Problem
〇Shu Nagasawa1, Zenji Yatabe1, Seiya Kasai1 (1. RCIQE, Hokkaido Univ.)
Keywords:
Analog calculations,Slime mold,Optimization problems,Electronic circuits
The analog electronic amoeba is a computer that electronically emulates the optimization capabilities of biological slime molds, and it is expected to efficiently search for solutions to combinatorial optimization problems. When solving the Boolean satisfiability problem (SAT) using conventional electronic amoebas, the pseudopods representing state variables could only be controlled for contraction (1 to 0). Consequently, two pseudopod units had to be assigned to represent a single variable. To prevent the increase in circuit scale caused by this redundant representation, this study designed a bidirectionally controllable pseudopod unit and attempted to verify its functionality by searching for solutions to a small-scale SAT. For the SAT solution search, in addition to the BB signal controlling the 1 to 0 transition, a feedback signal controlling the 0 to 1 transition was introduced to control the pseudopod units. To verify the search capability, an amoeba core comprising four pseudopod units and a BB circuit were implemented in a circuit simulator. The simulation results showed that each pseudopod unit transitioned from its initial state to reach the solution, demonstrating that the electronic amoeba equipped with the newly designed bidirectionally controllable pseudopod units is capable of solving the problem.