表 1. Statistical results of navigation experiments of various algorithms.

各算法导航实验的统计结果

环境	算法	导航路径 平均长度/m	发现目标区域后 收敛的概率	首次发现目标所需的 平均探索次数	完成导航习惯养成 所需的平均探索次数
1	Q-learning	16.59	84.2%	9.4	11.2
	SARSA	19.56	67.6%	10.7	16.5
	IAC	17.44	85.8%	11.9	17.3
	Sn-Plast	18.97	73.3%	6.6	13.8
	Sn-Plast + PO	13.02	73.3%	6.6	13.8
	Sn-Plast + PO + PF	12.48	96.1%	6.4	6.7
2	Q-learning	11.87	79.9%	8.5	9.3
	SARSA	12.62	73.5%	9.8	12.4
	IAC	11.14	81.3%	7.7	8.9
	Sn-Plast	12.90	80.7%	6.8	12.9
	Sn-Plast + PO	7.73	80.7%	6.8	12.9
	Sn-Plast + PO + PF	7.52	97.8%	7.0	7.1
3	Q-learning	16.28	76.2%	9.7	14.2
	SARSA	17.55	71.5%	9.3	22.8
	IAC	17.13	89.1%	8.5	16.1
	Sn-Plast	18.93	78.4%	7.9	20.7
	Sn-Plast + PO	11.40	78.4%	7.9	20.7
	Sn-Plast + PO + PF	11.16	97.3%	7.9	8.4
4	Q-learning	18.51	85.7%	9.7	14.1
	SARSA	19.38	70.2%	11.6	20.9
	IAC	18.96	91.5%	8.1	9.6
	Sn-Plast	19.73	74.8%	6.2	17.4
	Sn-Plast + PO	14.06	74.8%	6.2	17.4
	Sn-Plast + PO + PF	13.70	98.2%	5.9	6.1