SEIRモデルとMASモデルによるパンデミックシミュレーションの整合性

摘要

感染病蔓延の予測モデルには常微分方程式によるSEIRモデルとMAS（マルチエージェントシミュレーション）がある．SEIRモデルでは大人数のシミュレーションが短時間で可能であるが細かい条件の設定ができない．MASでは詳細なシミュレーションが可能であるが大人数では計算量は指数的に膨大になる．そこで，初期の段階でMASによる結果を条件としてSEIRに引き継ぐシミュレーション法を考えた．ここではその妥当性を確認するため，MASで通した計算結果と両者を組み合わせた計算結果との整合性を調べてみた．%There are mainly two kinds of methods for pandemic simulations: one is the SEIR model and the other is the MAS model, the multi-agent simulation. The SEIR model can deal with simulations quickly for many homogeneous populations with simple ordinary differential equations; however, we cannot set many detailed conditions to the model. The MAS model can deal with detailed simulations under the many kinds of initial and boundary conditions with simple social network models; however, the computing cost will be exponentially growing up as the population size becomes larger, thus simulations in big cities would hardly be realized unless supercomputers are available. Therefore, we consider a method to combine these two methods such that we use the MAS model in the early stage in a pandemic simulation to determine the appropriate parameters used in the SEIR model and then we use the MAS model after that with these obtained parameters. To investigate the validity of this combined method we first compare the simulation results between the SEIR model and the MAS model. In this paper we show the consistency of the results; provided the results, we will be able to use the combination method for many situations.