OK
Researcher profile

Okeke Ikenna Stephen

Academic / Lecturer · David Umahi Federal University of Health Sciences

Mathematics, Numerical Analysis and OptimizationIndustrial Mathematics and Health Statistics · Nigeria
6Linked publications
0Citations
0h-index
0i10-index

Metrics are calculated from publications currently linked to this profile.

Researcher overview

About

Okeke Ikenna Stephen is a KnowledgeTrend researcher working in Mathematics, Numerical Analysis and Optimization.

Research output

Recent Publications

6 research works linked to this profile

▤
Research · 2026 · Ktrend - International Journal of Computational Mathematics and Scientific Computing

Exact regret envelopes for commitment-constrained manufacturing under joint capacity stress

A feasible manufacturing adjustment can preserve mandatory commitments while losing most of the contribution attainable through reoptimization. This paper studies nominal-surplus scaling under simultaneous capacity losses and increasing production floors. For arbitrary finite product and resource counts with nonnegative resource use, it derives the exact feasible stress interval and a finite breakpoint audit of the largest scaled-plan regret. A two-product family with strictly positive resource coefficients and positive commitments proves that the optimum-to-scaled-value ratio can grow without bound. A dual-certificate tracing procedure supplies an interval-wide regret bracket without enumerating every dual vertex. The generic tracing principle is established parametric optimization; its specialization here couples commitment-adjusted scaling with a complete regret audit. An explicit two-resource model also gives exact product-mix transitions, shadow prices and coefficient-uncertainty limits. Experiments cover 90 manufacturing instances and compare oracle tracing, direct optimizer grids and independently enumerated dual envelopes. The trace requires a mean of 3.58 LP queries on these instances; 62 have an interior regret maximum, and a 101-point grid understates the largest regret by up to 0.4425 accounting units. These synthetic results support complete-path auditing rather than extrapolation from a nominal shadow price.

▤
Research · 2026 · Ktrend - International Journal of Mathematics and Statistics (IJMS)

Perturbation-stable failure and weighted dual repair for harmonic-mean transportation allocations

Harmonic-mean transportation rules can generate feasible schedules without providing a cost guarantee. This paper studies the rule that selects the active row or column of largest harmonic mean and allocates to its cheapest cell. An explicit positive-cost family has unbounded approximation ratio, and a perturbed family shows that the failure persists throughout a full-dimensional neighbourhood with strict deciding choices. For every route-cost perturbation of magnitude at most 0.01, the ratio is bounded below by \((M+11.97)/13.03\) for \(M\geq32\). A weighted potential-correction problem then computes the strongest transportation dual bound obtainable by downward movement of supplied candidate potentials. Its bipartite packing dual supplies an independent repair certificate, and the resulting bound dominates a scalar correction. Computation includes 300 assignment instances, 600 perturbation tests and 120 rectangular transportation instances. The weighted correction improves the scalar bound on all rectangular cases, with a mean correction-loss ratio of approximately 0.372. The worst-case theorem concerns the specified initial rule; it does not apply to subsequent exact improvement. The study combines a perturbation-stable negative result with an auditable certificate framework. All numerical data are synthetic, and the underlying LP duality is established.

▤
Caputo derivative; HIV dynamics; CD4+ T cells; local stability; hybrid activation; bounded neural modifier. · 2026 · African Journal of Mathematics, Statistics and Computer Science

Stability Thresholds and Bounded Hybrid Neural Modifiers in a Fractional HIV–CD4+ T-Cell Model

A commensurate Caputo model of HIV infection of CD4+ T cells is analysed by separating biologically admissible equilibria from algebraic roots. Building on the fractional model of Okeke et al. [1] and the hybrid activation construction of Essang et al. [2], this study introduces bounded neural modifiers of infection and viral production rates. Explicit equilibrium expressions and an infection threshold are derived. The disease-free equilibrium is locally asymptotically stable below the threshold, while a unique positive endemic equilibrium exists and is locally asymptotically stable above it for every common order in $(0,1]$. A recalculation using the parameter values reported in [1] produces a stable endemic equilibrium, contrary to the published numerical classification. A conservative bound on neural modifiers provides a sufficient condition for local disease-free stability across fixed covariate settings. A completed numerical study reports dimensionless fractional trajectories, threshold surfaces, Jacobian spectra and time-step refinement. These results concern mathematical equilibria rather than clinical treatment efficacy.

▤
structural identifiability; fractional HIV model; hybrid activation; inverse problem; slope bounds; physics-informed neural networks · 2026 · Ktrend – Nigerian Journal of Mathematical and Computational Sciences

Structural Identifiability and Gradient Bounds for Hybrid-Activation Learning of Fractional HIV Dynamics

The recovery of biological parameters from fractional HIV trajectories is examined alongside the mathematical properties of hybrid neural activations. Using the model of Okeke et al. [1] and the five-component activation mixture of Essang et al. [2], an exact parameter symmetry is derived: proliferation rate, healthy-cell death rate and carrying capacity cannot all be recovered separately from the model trajectories without additional information. A reduced parametrisation eliminates this redundancy. Conditional identifiability of the reduced coefficients is established through full-rank trajectory regressors for a known fractional order and fully observed states. The hybrid activation derivative is corrected, and global scalar slope bounds are proved. These bounds do not guarantee preservation of gradients through arbitrary network depth. A fractional residual learning framework is specified, and completed synthetic experiments examine parameter symmetry, regressor singular values, observation-window conditioning and slope products. The study establishes mathematical and computational limitations rather than empirical superiority of a trained learning algorithm.

▤
Algebra · 2026 · International Journal of Mathematics and Statistics

Differential Semigroups of Automaton Perturbations and Incremental Syntactic Reconstruction

This paper investigates the enumeration of elements in three classes of signed partial transformation semigroups: partial signed order-preserving, partial signed order-decreasing, and partial signed order-preserving-or-order-reversing transformations. It examines the reported PSPD₄ numerical inconsistency and develops a domain-rank framework for extending the finite enumerations, while distinguishing interpolated computational conjectures from exact counting formulae.

▤
Research · 2026 · KnowledgeTrend Journal

Certified bottleneck transitions and balanced capacity upgrades in robust production–distribution networks

Manufacturing resources and shipping routes can constrain the same customer commitment, making both uncertainty audits and capacity spending sensitive to changing bottlenecks. This paper develops a certified production–distribution analysis under a static uncertainty model with common resource-coefficient inflation. A reparameterization converts the uncertainty path to a fixed-matrix LP and yields a finite dual-line envelope for maximal supportable commitment stress. A certificate-driven interval procedure traces transitions with explicit upper and lower bounds. A budgeted resource–route upgrade model then derives marginal budget value by minimizing the largest price-normalized multiplier over the entire baseline-optimal dual face. A family with \(k\) simultaneous bottlenecks shows that an arbitrary extreme certificate can overstate marginal value by a factor \(k\), while upgrading only its indicated bottleneck produces zero gain. An explicit two-plant family provides an exact switching law and constructive shipment proof. Computation covers 36 networks with up to 144 shipment variables, 180 budgeted upgrade cases and 25 serial-bottleneck cases. Independently solved primal and dual programs verify the envelope and spending calculations. The method specializes established LP and parametric-optimization principles; its contribution is the linked bottleneck-and-upgrade audit with explicit failure families. All experiments are synthetic.

ViewFull file not uploaded