--- title: Robust Downhill Simplex Method (rDSM) url: https://www.emergentmind.com/topics/robust-downhill-simplex-method-rdsm type: topic --- # Robust Downhill Simplex Method (rDSM) Robust Downhill Simplex Method (rDSM) denotes a robustness-oriented variant of the Downhill Simplex Method (DSM), or Nelder–Mead family, for derivative-free minimization, introduced as a software package for optimization problems in high dimensions and in the presence of noise [2509.05917]. It preserves the standard DSM geometric moves—reflection, expansion, contraction, and shrink—but augments them with two explicit mechanisms: simplex degeneracy detection and correction, and reevaluation of long-standing vertices to estimate the real objective value of noisy problems. In the broader simplex literature, closely related robustness ideas appear as random restart when simplex points “get near each other” [1209.0853], regular-simplex geometry combined with sufficient decrease and shrinking rules that admit worst-case complexity bounds [2508.16561], and hybrids that alternate a downhill simplex step with explorative Latin hypercube sampling [1905.12036]. Taken together, these works place rDSM within a larger program of making simplex-based search less sensitive to premature convergence, degenerated simplices, and noise-induced spurious minima. ## 1. Classical descent framework and rDSM problem setting rDSM solves the unconstrained derivative-free minimization problem \[ \text{Find } x^\star \in \mathbb{R}^n \text{ that minimizes } J(x), \] where no gradient \(\nabla J\) is assumed available and evaluation of \(J\) may be noisy [2509.05917]. As in classical DSM, the current state is an \(n\)-simplex \[ S = \{x^{s_1},\dots,x^{s_{n+1}}\}, \] with vertices ordered so that \[ J(x^{s_1}) \le \dots \le J(x^{s_n}) \le J(x^{s_{n+1}}), \] where \(x^{s_1}\) is best and \(x^{s_{n+1}}\) is worst. The centroid of all but the worst point is \[ x^{c} = \frac{1}{n} \sum_{i=1}^{n} x^{s_i}. \] The underlying DSM moves retained by rDSM are the standard Nelder–Mead operations. Reflection uses \[ x^{r} = x^{c} + \alpha (x^{c} - x^{s_{n+1}}), \] expansion uses \[ x^{e} = x^{c} + \gamma (x^{c} - x^{s_{n+1}}), \] outside and inside contraction use \[ x^{co} = x^{c} + \rho (x^{c} - x^{s_{n+1}}), \qquad x^{ci} = x^{c} - \rho (x^{c} - x^{s_{n+1}}), \] and shrink updates \[ x^{s_i} \leftarrow x^{s_1} + \sigma (x^{s_i} - x^{s_1}), \qquad 0<\sigma<1. \] The rDSM software uses the standard Nelder–Mead parameters \[ \alpha = 1,\qquad \gamma = 2,\qquad \rho = 0.5,\qquad \sigma = 0.5. \] The key structural point is that rDSM does not alter the basic DSM decision logic for reflection, expansion, contraction, and shrink. Instead, it wraps that loop with corrective procedures. The main iteration is: run a Nelder–Mead iteration, check degeneracy, call degeneracy correction if required, apply reevaluation logic for long-standing vertices, and proceed to the next iteration [2509.05917]. A common misconception is therefore that rDSM is a wholly different simplex optimizer. In the 2025 formulation, it is more precisely a standard DSM augmented by robustness mechanisms around the basic geometric moves. ## 2. Degeneracy detection and geometric reinflation The first robustness mechanism addresses simplex degeneration, described as a major source of loss of effective dimensionality, poor exploration of the search space, and stagnation or premature convergence at non-stationary points [2509.05917]. The method distinguishes two types of degeneracy: an **edge-degenerated simplex**, in which one edge is much shorter than others, and a **volume-degenerated simplex**, in which the volume is too small relative to the edge lengths. Because each type may occur without the other, rDSM uses two independent criteria. The simplex perimeter is defined as \[ P(x^{s_1},\dots,x^{s_{n+1}}) = \sum_{i=1}^{n} \sum_{j=i+1}^{n+1} \left\|x^{s_i} - x^{s_j}\right\|, \] and the \(n\)-dimensional volume is \[ V(x^{s_1},\ldots,x^{s_{n+1}}) = \frac{1}{n!}\left| \det \begin{bmatrix} x^{s_1} & x^{s_2} & \dots & x^{s_{n+1}}\\ 1 & 1 & \dots & 1 \end{bmatrix} \right|. \] For edge-based degeneracy, with edge vectors \(\mathbf{e}^i\) and norms \(\|\mathbf{e}^i\|\), \[ \varepsilon_e = \frac{\min_i \|\mathbf{e}^i\|}{\max_i \|\mathbf{e}^i\|}, \] and the simplex is declared edge-degenerated if \[ \varepsilon_e < \theta_e. \] For volume-based degeneracy, \[ \varepsilon_v = \sqrt[n]{ \frac{\det[\mathbf{e}]}{\prod_i \|\mathbf{e}^i\|} }, \] and the simplex is volume-degenerated if \[ \varepsilon_v < \theta_v. \] The default thresholds are \(\theta_e = 0.1\) and \(\theta_v = 0.1\). Once degeneracy is detected, rDSM repairs the simplex by moving the worst vertex \(x^{s_{n+1}}\) to a new position \(y^{s_{n+1}}\) that maximizes simplex volume while preserving simplex perimeter: \[ \begin{cases} \displaystyle y^{s_{n+1}} = \arg\max_{\mathbf{y}} \; V(x^{s_1},\dots,x^{s_n}, \mathbf{y})\\[0.7em] \text{subject to } P(x^{s_1},\dots,x^{s_n}, \mathbf{y}) = P(x^{s_1},\dots,x^{s_n}, x^{s_{n+1}}). \end{cases} \] This constrained problem is solved by a Newton–Raphson method. The perimeter constraint keeps the overall scale of the simplex similar, while the volume objective inflates the simplex in collapsed directions. If moving the worst vertex does not remove degeneracy, the algorithm proceeds to move the next worst vertex, and so on, until the simplex is nondegenerate or all points have been corrected [2509.05917]. In two dimensions, the geometric interpretation is explicit: the locus of points with constant sum of distances to two fixed vertices is an ellipse, so the corrected vertex lies on an ellipse with foci at the fixed vertices, and the algorithm selects the location giving maximal area [2509.05917]. This suggests a useful conceptual reading of rDSM: rather than restarting the search globally, it performs an internal reinitialization that attempts to preserve accumulated information while restoring full-dimensional search geometry. ## 3. Noise-robust objective estimation by reevaluation The second robustness mechanism addresses noisy objectives. The paper evaluates rDSM under additive noise, \[ \tilde{J}(x) = J(x) + \eta, \] with uniform noise \(\mathcal{U}[0,b]\), where \(b = 0.01\) or \(0.02\), and Gaussian noise \(\mathcal{N}(0,\sigma^2)\), where \(\sigma^2 = 0.005\) or \(0.01\) [2509.05917]. The stated concern is that classical DSM reacts to whichever noisy value it sees and can be misled into local, noise-induced minima. To mitigate this, rDSM assigns to each vertex \(x^{s_i}\) a counter \(c^{s_i}\) that records how many DSM iterations the vertex has remained in the current simplex. The key observation is that most vertices leave the simplex after roughly \(n\) iterations; a vertex that remains significantly longer is likely close to a search center where the algorithm is oscillating. If \[ c^{s_i} \ge 1.5\,n, \] the algorithm reevaluates \(J(x^{s_i})\) and replaces the stored objective value by the average of all evaluations at that point: \[ \hat{J}(x^{s_i}) = \frac{1}{m} \sum_{k=1}^{m} J_k(x^{s_i}), \] where \(J_k(x^{s_i})\) are the \(m\) evaluations performed at \(x^{s_i}\) so far. The averaging is intended to reduce the impact of unbiased noise by lowering the variance of the estimate. This reevaluation step affects the ranking of simplex vertices, and therefore the reflection direction and subsequent move acceptance. In the reported tests, the threshold \(1.5n\) was preferred over a more conservative \(2n\): using \(2n\) doubled average runtime from \(4.4\text{ s}\) to \(8.3\text{ s}\) on a representative 2D noisy test and yielded worse accuracy, converging to a point far from the true optimum [2509.05917]. The mechanism is deliberately selective rather than universal; the method does not reevaluate every point, but concentrates repeated measurement on long-standing vertices that disproportionately influence the local search trajectory. ## 4. Software implementation and empirical behavior The rDSM software package is implemented in MATLAB, tested with MATLAB 2021b on Microsoft Windows, and released at `https://github.com/tianyubobo/rDSM` under a CC-BY-SA license [2509.05917]. Its directory structure separates `Main`, `ObjectiveFunction`, `Initialization`, `Optimizer`, `Visualization`, and `Output`. The optimizer directory contains `DSM.m`, described as a reimplementation of `fminsearch`, and `rDSM.m`, which adds degeneracy correction and reevaluation. The package writes histories such as `SimplexHistory.txt`, `PointsDatabase.txt`, and `ReevaluationHistory.txt`, together with `.dat` ASCII versions. The reported validation emphasizes three settings. In a 2D linear-gradient problem without an obstacle, \[ J(x_1,x_2) = -\frac{x_1 - x_2}{4} + 0.5,\qquad (x_1,x_2)\in[-1,1]^2, \] DSM starting from \((-0.75, 0.35)\) converges rapidly to \((0.9999, -0.9998)\) with \(J \approx 9.21\times 10^{-5}\), essentially the global minimum. With an obstacle defined by \(J=10^3\) in the region \(-1