Jaya R Package -- A Parameter-Free Solution for Advanced Single and Multi-Objective Optimization

Neeraj Dhanraj Bokde

REVIEW 4 major objections 5 minor 14 references

Jaya R Package -- A Parameter-Free Solution for Advanced Single and Multi-Objective Optimization

T0 review · 4 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The Jaya R package claims parameter-free single- and multi-objective optimization.

desk verdict Reasonable R package, but multi-objective claims outrun the evidence; benchmark needs statistical rigor. read the letter →

arxiv 2411.16509 v1 pith:HYUUJVTV submitted 2024-11-25 cs.MS cs.LG

Neeraj Dhanraj Bokde This is my paper · ORCID

classification cs.MScs.LG

keywords Jayaalgorithmparameter-freeoptimizationmulti-objectiveParetofrontRpackageconstrainthandlingbenchmarkfunctionsrenewableenergy

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper introduces an R package that implements the Jaya algorithm, a population-based optimizer with no algorithm-specific parameters. It claims the package solves single-objective problems and, through Pareto front tracking, multi-objective problems, while offering constraint handling, adaptive population sizing, early stopping, and parallel evaluation. The authors demonstrate the package on standard benchmark functions and on a renewable-energy mix case study. If the claims hold, users get a tuning-free optimizer that returns a set of trade-off solutions for conflicting objectives.

What carries the argument

The load-bearing mechanism is the update rule $$x_{\text{new}} = x_{\text{old}} + r_1 \cdot (x_{\text{best}} - |x_{\text{old}}|) - r_2 \cdot (x_{\text{worst}} - |x_{\text{old}}|)$$ with $r_1, r_2$ uniform in $[0,1]$. The rule requires no tuning parameters. In the multi-objective variant, the package tracks the set of non-dominated solutions, which becomes the reported Pareto front.

What would settle it

Run each benchmark problem (for example, 30 independent runs) with Jaya and with DE, GA, SA, PSO, and NM using their standard recommended settings, then compare mean best objective values and their variances; if Jaya is not statistically competitive on Rosenbrock or Ackley, the reported performance claim fails. For the multi-objective claim, compare jaya_multi()'s Pareto front against a known reference front on a standard test problem such as ZDT1 or DTLZ1.

Watch

Extended reading notes

Core claim

The central claim is that the Jaya update rule, which iteratively moves each candidate toward the current best solution and away from the current worst solution using two uniform random coefficients, is sufficient to handle a wide range of constrained and multi-objective problems. The paper packages this rule into the jaya() and jaya_multi() functions, where jaya_multi() maintains a Pareto front of non-dominated solutions and returns them together with their decision variables and objective values.

Load-bearing premise

The benchmark comparison assumes that a single run of each algorithm, with unspecified parameter settings and no statistical analysis, is a fair and representative measure of performance.

Editorial extensions

If this is right

Users can run optimization in R without setting crossover rates, mutation probabilities, inertia weights, or other algorithm-specific hyperparameters.
Multi-objective problems return a Pareto front, letting decision-makers choose among trade-off solutions rather than a single compromise.
The renewable-energy case study suggests the package can balance carbon emissions, cost, and reliability in a policy-relevant energy mix problem.
Parallel evaluation and early stopping are claimed to reduce computation time on large-scale problems.
Benchmark results on Sphere, Rastrigin, Rosenbrock, Ackley, and Griewank support the claim that Jaya is competitive with DE, GA, SA, PSO, and Nelder-Mead.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

A natural next test is to compare jaya_multi() against established multi-objective algorithms such as NSGA-II or SPEA2 using statistical replicates and standard performance indicators; the paper's single-run benchmark does not establish dominance.
The 'parameter-free' claim applies to the core update rule, but users must still choose population size, iteration count, and convergence tolerance, so the practical tuning burden is reduced rather than eliminated.
The energy case study could be extended with real cost and capacity data to see whether the reported Pareto front persists under actual engineering constraints and stochastic supply profiles.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Quick take: this is a plausible software paper with a real R package, but the multi-objective claims outrun the evidence, and the benchmark section needs statistical work before it can support the abstract's language.

The genuine contribution is the package itself: a parameter-free Jaya implementation in R with constraint handling, adaptive population, early stopping, parallel evaluation, and plotting helpers. The single-objective example is clean and reproducible, and the author points to public code for both the benchmarks and the energy case study. That is the right way to write a software paper, and the package looks like it could be useful to R users who want a tuning-free optimizer.

The soft spots are where the paper starts claiming more than it shows. The multi-objective section is the weakest. The paper says jaya_multi does 'Pareto front tracking' but never states how the best/worst update rule from Eq. (1) is generalized when there is no unique xbest or xworst. No archive, no dominance relation, no selection rule is given. The example shows a scatter plot of two sphere objectives, but a reader cannot tell whether the output is a genuine Pareto front or just a cloud of evaluated solutions. The same gap appears in the energy case study: it plots 3D fronts and discusses wind, solar, and hydro trade-offs, but gives no objective functions, no equations, no data, and no metrics (hypervolume, IGD, etc.) to quantify front quality. The stress-test note is right: the core multi-objective claim is unsubstantiated in the manuscript.

The benchmark table for single-objective problems is also thin. There are no repeated runs, no error bars, no explicit statement of dimensions or bounds, and no parameter settings for the comparators (DE, GA, SA, PSO, NM). Single runs can look nice by luck. That said, the Jaya numbers on Sphere and Rastrigin are so small that the qualitative point—Jaya is competitive—probably survives a proper comparison. This is a moderate weakness, not a fatal one.

No circularity issues: the self-citations are to the author's package and data repo, which is normal for software papers. The paper does not try to pass fitted values off as predictions.

The bottom line: the single-objective package is a legitimate, modest contribution. The multi-objective capability may exist in the code, but the paper does not show it. I would send this to peer review with clear requests: add the multi-objective algorithm description, provide a statistical benchmark protocol, and either add equations/data for the energy case study or cut it down to a brief illustration. A serious referee can fix this; the package itself seems real.

Referee Report

4 major / 5 minor

Summary. The paper presents the R package Jaya, an implementation of the parameter-free Jaya optimization algorithm. It describes the jaya() function for single-objective optimization and the jaya_multi() function for multi-objective problems, and it discusses features such as constraint handling, adaptive population size, early stopping, and parallel processing. The package is demonstrated with a sphere-function example, a two-objective example, a five-benchmark comparison table, and a qualitative energy-system case study. The central claim is that the package provides a robust and versatile parameter-free optimizer for both single- and multi-objective optimization problems.

Significance. If the claims are substantiated, the package would be a useful addition to the R optimization ecosystem, because Jaya has no algorithm-specific control parameters and the package provides convenience functions for summaries and plots. The public code repository, the runnable examples, and the attempt at benchmark comparisons are concrete strengths. However, the multi-objective capability is not algorithmically specified, and the performance comparisons are not quantitatively validated, so the current evidence does not yet support the stated scope of the package.

major comments (4)

[Multi-Objective Optimization (near Fig. 3)] The manuscript never specifies how jaya_multi() generalizes the Jaya update rule to multiple objectives. Equation (1) requires a unique xbest and xworst, but in multi-objective optimization there is generally no single best and worst solution. The text only says that 'Pareto front tracking' is used; it does not describe how non-dominated solutions are selected, how an archive is maintained, how dominance is defined, or how the best/worst solutions are chosen for the update. As a result, the Pareto_Front output cannot be distinguished from an arbitrary set of evaluated points. Please provide the exact algorithmic variant implemented, including a formal description or pseudocode, and validate the output on a problem with a known Pareto front.
[Table 1 and 'Performance Comparison of Optimization Algorithms'] The benchmark comparison reports only single achieved values for each algorithm. It does not state the dimension n, the bounds of the search space, the population sizes, the number of iterations, the convergence tolerances, or the number of independent runs for Jaya and each comparator. Without this experimental protocol and without repeated-run statistics, the claim of competitive performance is unsupported. In addition, exact zeros reported for SA, PSO, and NM on several benchmarks require explanation (for example, analytical solutions or tight tolerances). Please report the full setup, provide mean and standard deviation over multiple runs, and include the script used to generate the table.
[Application of the Jaya Package in Multi-Objective Energy Optimization] The energy case study is qualitative only. The objectives (carbon emissions, total costs, reliability) and the constraints (renewable contribution above 70%, variable bounds of 10-40%) are described in words, but no mathematical definitions are given for the objective functions, the reliability index, or the cost models. No numerical results are reported, and no comparison is made with other multi-objective optimizers or with a known baseline. The text around Figures 4-7 draws conclusions about wind, solar, and hydro contributions, yet the figures are not embedded and no quantitative summary of the Pareto front is provided. Please add the complete problem formulation, data sources, simulation parameters, quantitative results, and reproducibility instructions.
[Overall claim in the abstract] The abstract and title state that the package is 'parameter-free' and 'suitable for solving both single-objective and multi-objective optimization problems.' The multi-objective part of this claim is not supported by the algorithmic description, and the single-objective performance claim is not supported by the under-specified benchmark. These are not merely presentation issues; they affect the central contribution of the paper and should be corrected before publication.

minor comments (5)

[Title and Section 'Overview of the Jaya R Package'] The phrase 'parameter-free' is presented absolutely, while the package API includes popSize, maxiter, tolerance, patience, and an adaptive population range. Consider clarifying that the algorithm has no algorithm-specific parameters, while the implementation still requires user-specified population and termination settings.
[Equation (1)] Equation (1) uses |xold| without stating that the absolute value is applied componentwise to the decision-variable vector; adding this clarification would avoid ambiguity for readers implementing the update.
[Table 1] The benchmark equations are typeset inside a single wide table cell with line breaks; splitting the equations into separate rows or using display math would improve readability.
[Reproducibility] The manuscript cites GitHub repositories for the package and case-study code, but does not provide a versioned archive or DOI. A persistent archival reference, such as a CRAN submission or Zenodo DOI, would improve reproducibility for readers.
[Figures 4-7] Figures 4-7 are referenced with captions but the figures do not appear in the manuscript text. If this is a formatting issue, please ensure all figures are embedded; otherwise the case-study discussion is difficult to follow.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity; the paper is a software description with empirical benchmark results, and its self-citations are provenance pointers rather than load-bearing derivational evidence.

full rationale

This is a software-description paper with no derivation chain whose conclusions could reduce to inputs by construction. The Jaya update rule in Section 2 is stated as a standard population-based heuristic, and the benchmark outcomes in Table 1 are empirical optimization results, not quantities fitted from those results. No hyperparameters are fitted to the benchmarks, and no prediction is generated from a fitted model. The self-citations, to the package repository (Bokde) and the case-study code (Bokde, 2024), are provenance pointers rather than load-bearing evidence for any mathematical claim. The main weaknesses are evidentiary, not circular: the multi-objective section never presents the algorithm by which jaya_multi handles dominance, non-dominance, or archive maintenance, and the single-run benchmark table in Table 1 lacks statistical analysis. Those concerns affect whether the claims are substantiated, but they are not circularity. Score 1 reflects only the presence of routine self-citations in an otherwise self-contained software report.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

The central claim relies on the correctness and convergence of the Jaya algorithm and on the fairness of the benchmark comparison. No new theoretical entities are introduced.

assumptions (3)

domain assumption The Jaya algorithm's update rule converges to optimal or near-optimal solutions for the tested benchmark problems.
Stated in Section 2 and used in all experiments, but no convergence proof is provided.
standard math The benchmark function definitions match the cited literature and are evaluated correctly.
The paper lists the equations but does not provide verification code within the manuscript.
domain assumption The comparison algorithms (DE, GA, SA, PSO, NM) were configured appropriately and their results are representative.
The paper does not specify the parameters used for these algorithms, so fairness of the comparison is assumed.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Jaya R Package -- A Parameter-Free Solution for Advanced Single and Multi-Objective Optimization." pith.science (2026). https://pith.science/paper/HYUUJVTV

@misc{pith2026241116509,
  author       = {Pith},
  title        = {Pith review of: Jaya R Package -- A Parameter-Free Solution for Advanced Single and Multi-Objective Optimization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HYUUJVTV}},
  note         = {Machine review of arXiv:2411.16509}
}

read the original abstract

The Jaya R package offers a robust and versatile implementation of the parameter-free Jaya optimization algorithm, suitable for solving both single-objective and multi-objective optimization problems. By integrating advanced features such as constraint handling, adaptive population management, Pareto front tracking for multi-objective trade-offs, and parallel processing for computational efficiency, the package caters to a wide range of optimization challenges. Its intuitive design and flexibility allow users to solve complex, real-world problems across various domains. To demonstrate its practical utility, a case study on energy modeling explores the optimization of renewable energy shares, showcasing the package's ability to minimize carbon emissions and costs while enhancing system reliability. The Jaya R package is an invaluable tool for researchers and practitioners seeking efficient and adaptive optimization solutions.

Figures

Figures reproduced from arXiv: 2411.16509 by the authors.

**Figure 1.** Block diagram of the Jaya algorithm. The core of the algorithm lies in its update rule. Each candidate solution is adjusted based on its proximity to the best and worst solutions in the current population. This adjustment is governed by the following equation: xnew = xold + r1 · (xbest − |xold|) − r2 · (xworst − |xold|) Here, xnew represents the updated solution, xold is the current solution, xbest is the best solut… view at source ↗

**Figure 2.** Convergence plot showing the evolution of the best objective value over iterations for minimizing a sphere function using the Jaya algorithm. • Objective Function: The Sphere function, f(x) = ∑ n i=1 x 2 i , is used as the objective function. It evaluates how close the variables are to zero, with the global minimum at f(x) = 0, when xi = 0 for all i. • Bounds: Decision variables are bounded between -5 and 5, ensurin… view at source ↗

**Figure 3.** Pairwise scatter plots of the Pareto front obtained from the multi-objective optimization of two sphere functions. The plots show trade-offs between objectives: minimizing the sum of squares and minimizing the sum of squares offset by 2. The plot_jaya_multi_pairwise(result_multi) function generates pairwise scatter plots ( [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗

Figures from the paper (4 more)

**Figure 4.** Figure 4: 3D Pareto Front of Energy System Optimization (Colored by Total Contribution) [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]

**Figure 5.** Figure 5: 3D Pareto Front of Energy System Optimization (Colored by Wind Contribution). Preprint submitted to the R Journal [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]

**Figure 6.** Figure 6: 3D Pareto Front of Energy System Optimization (Colored by Solar Contribution) [PITH_FULL_IMAGE:figures/full_fig_p011_6.png]

**Figure 7.** Figure 7: 3D Pareto Front of Energy System Optimization (Colored by Hydro Contribution). Preprint submitted to the R Journal [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

14 extracted references · 13 canonical work pages

[1]

D. Ackley. A connectionist machine for genetic hillclimbing, volume 28. Springer science & business media, 2012

work page 2012
[2]

Ardia, K

D. Ardia, K. Boudt, P. Carl, K. Mullen, and B. G. Peterson. Differential evolution with deoptim: an application to non-convex portfolio optimization. The R Journal, 3 0 (1): 0 27--34, 2011

work page 2011
[3]

Bendtsen

C. Bendtsen. pso: Particle Swarm Optimization, 2022. URL https://CRAN.R-project.org/package=pso. R package version 1.0.4

work page 2022
[4]

N. Bokde. Jaya: Gradient-Free Optimization Algorithm for Single and Multi-Objective Problems. URL https://github.com/neerajdhanraj/Jaya. R package version 1.0.1

work page
[5]

N. D. Bokde. Jaya case studies and comparisons. https://github.com/neerajdhanraj/Jaya_Case_Studies, 2024. Accessed: 2024-11-24

work page 2024
[6]

de JONG Kenneth

A. de JONG Kenneth. Analysis of the behavior of a class of genetic adaptive systems. Technical Report No. 185, Department of Computer and Communication Sciences, University of Michigan, 1975

work page 1975
[7]

A. O. Griewank. Generalized descent for global optimization. Journal of optimization theory and applications, 34: 0 11--39, 1981

work page 1981
[8]

Mullen, D

K. Mullen, D. Ardia, D. L. Gil, D. Windover, and J. Cline. Deoptim: An r package for global optimization by differential evolution. Journal of Statistical Software, 40 0 (6): 0 1--26, 2011

work page 2011

Show all 14 references

[9]

R: A Language and Environment for Statistical Computing

R Core Team . R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria, 2024. URL https://www.R-project.org/

2024
[10]

R. Rao. Jaya: A simple and new optimization algorithm for solving constrained and unconstrained optimization problems. International Journal of Industrial Engineering Computations, 7 0 (1): 0 19--34, 2016

2016
[11]

L. A. Rastrigin. Systems of extremal control. Nauka, 1974

1974
[12]

Rosenbrock

H. Rosenbrock. An automatic method for finding the greatest or least value of a function. The computer journal, 3 0 (3): 0 175--184, 1960

1960
[13]

L. Scrucca. Ga: A package for genetic algorithms in r. Journal of Statistical Software, 53: 0 1--37, 2013

2013
[14]

L. Scrucca. On Some Extensions to GA Package: Hybrid Optimisation, Parallelisation and Islands Evolution . The R Journal , 9 0 (1): 0 187--206, 2017. doi:10.32614/RJ-2017-008. URL https://doi.org/10.32614/RJ-2017-008

2017 doi

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [1]

D. Ackley. A connectionist machine for genetic hillclimbing, volume 28. Springer science & business media, 2012

work page 2012

[2] [2]

Ardia, K

D. Ardia, K. Boudt, P. Carl, K. Mullen, and B. G. Peterson. Differential evolution with deoptim: an application to non-convex portfolio optimization. The R Journal, 3 0 (1): 0 27--34, 2011

work page 2011

[3] [3]

Bendtsen

C. Bendtsen. pso: Particle Swarm Optimization, 2022. URL https://CRAN.R-project.org/package=pso. R package version 1.0.4

work page 2022

[4] [4]

N. Bokde. Jaya: Gradient-Free Optimization Algorithm for Single and Multi-Objective Problems. URL https://github.com/neerajdhanraj/Jaya. R package version 1.0.1

work page

[5] [5]

N. D. Bokde. Jaya case studies and comparisons. https://github.com/neerajdhanraj/Jaya_Case_Studies, 2024. Accessed: 2024-11-24

work page 2024

[6] [6]

de JONG Kenneth

A. de JONG Kenneth. Analysis of the behavior of a class of genetic adaptive systems. Technical Report No. 185, Department of Computer and Communication Sciences, University of Michigan, 1975

work page 1975

[7] [7]

A. O. Griewank. Generalized descent for global optimization. Journal of optimization theory and applications, 34: 0 11--39, 1981

work page 1981

[8] [8]

Mullen, D

K. Mullen, D. Ardia, D. L. Gil, D. Windover, and J. Cline. Deoptim: An r package for global optimization by differential evolution. Journal of Statistical Software, 40 0 (6): 0 1--26, 2011

work page 2011