2015 年 80 巻 711 号 p. 389-396
A scale-free power law distribution applies to fire-spreading networks in areas that are densely crowded with wooden-building, with each flammable wooden building as a node and its links to other buildings within fire-spreading distance of it. Based on the percolation theory of scale free networks, we have computed the percolation critical exponents and found that fire-spreading networks can be effectively disrupted by the targeted removal of each fire-spreading wooden building that links to six or more other such buildings. A targeted removal approach for the 84 highest-risk districts in Tokyo can prevent fire spread by removing 10-30% of flammable buildings, in contrast to random removal of 40-70% of flammable buildings. A detailed area survey and assortative coefficient indicate that cooperative rebuilding would be effective for a targeted removal approach.