In Quasi-Newton method, the matrix $left[B_{i} ight]$ is updated using the formula [left[B_{i+1} ight]=left[B_{i} ight]+frac{lambda_{i}^{*} s_{i} s_{i}^{T}}{s_{i}^{T} g_{i}}-frac{left(left[B_{i} ight] g_{i} ight)left(left[B_{i} ight] g_{i} ight)^{T}}{left(left[B_{i} ight]

In Quasi-Newton method, the matrix $\left[B_{i}\right]$ is updated using the formula

\[\left[B_{i+1}\right]=\left[B_{i}\right]+\frac{\lambda_{i}^{*} s_{i} s_{i}^{T}}{s_{i}^{T} g_{i}}-\frac{\left(\left[B_{i}\right] g_{i}\right)\left(\left[B_{i}\right] g_{i}\right)^{T}}{\left(\left[B_{i}\right] g_{i}\right)^{T} g_{i}}\]

Discuss the effect of $\lambda_{i}^{*}$ in this iteration process.