Algorithm 1 Needs learning algorithm

Initialize the independent variable of basic needs $x_{\mathrm{BNS}}$ and the one of growth needs  $x_{\mathrm{GNS}}$  Initialize the maximum exploration probability $r_{max}$  Initialize M tables of basic needs and growth needs  Initialize c = 1  Repeat (for each episode):   r ← $\frac{r_{max}}{2l-1}$ $x_{\mathrm{BNS}}$ ← $\frac{1}{1+e^{10\mathrm{c}{f}_{\mathrm{BNS}}\left(\mathrm{s}\right)}}$ $x_{\mathrm{GNS}}$ ← $\frac{1}{1+e^{10\mathrm{c}{f}_{\mathrm{GNS}}\left(\mathrm{s}\right)-5}}-x_{\mathrm{BNS}}$   c ← 1+ɑ$x_{\mathrm{GNS}}$   If $x_{\mathrm{GNS}}$>$x_{\mathrm{BNS}}$ then     Dominant Need ← growth needs   Else     Dominant Need ← basic needs   Choose $A_m$,$A_n$ from A at random   Choose $r_0$ randomly from the interval [0, 1)   If $r_0$< r then $A_k$ = {$A_m$,$A_n$}     Save $A_k$ into A   Else     If s in M table of dominant need then       Take $A_k$ with the largest M value from the M table of dominant need     Else $A_k$ ← $A_n$   Execute action $A_k$ to obtain the environmental improvement amount $M_s$   Update M table of dominant need with $A_k$, s, $M_s$