Rosetta Code: Knapsack Problem/Unbounded

[mlpfanatic903]

Hello forum,
I am currently investigating a particular problem in the Rosetta Code section, Knapsack Problem/Unbounded, and the particular solution provided to the Hint section in the worst-case scenario.

Below is the solution provided…

function knapsackUnbounded(items, maxweight, maxvolume) {
    var n = items.length;
    var best_value = 0;
    var count = new Array(n)
    var best = new Array(n)
    function recurseKnapsack(i, value, weight, volume) {
        var j, m1, m2, m;
        if (i == n) {
            if (value > best_value) {
                best_value = value;
                for (j = 0; j < n; j++) {
                    best[j] = count[j];
                }
            }
            return;
        }
        m1 = Math.floor(weight / items[i].weight);
        m2 = Math.floor(volume / items[i].volume);
        m = m1 < m2 ? m1 : m2;
        for (count[i] = m; count[i] >= 0; count[i]--) {
            recurseKnapsack(
                i + 1,
                value + count[i] * items[i].value,
                weight - count[i] * items[i].weight,
                volume - count[i] * items[i].volume
            );
        }
    }
    recurseKnapsack(0, 0, maxweight, maxvolume);
    return best_value;
}

console.log(`Case 4 best_val: ${knapsackUnbounded([{ name:"panacea", value:3000, weight:0.3, volume:0.025 }, { name:"ichor", value:1800, weight:0.2, volume:0.015 }, { name:"gold", value:2500, weight:2, volume:0.002 }], 35, 0.35)}\n`);

This is the supposed issue with the solution: a discrepancy between the solution and the presented value, of 600 (best_value_solution = 75300, best_value_presentedValue = 75900)…

Is there another way to resolve this discrepancy given the parameters (in JSON/ dictionary), and/or improve the proposed solution?

Thank you for your input

[sanity]

Solution wasn’t updated here when test was corrected. There’s a slight precision problem, that’s causing incorrect result. Take a look at the issue Rosetta Code: Knapsack problem/Unbounded - one incorrect test · Issue #39398 · freeCodeCamp/freeCodeCamp · GitHub

[JeremyLT]

I went ahead and update the solution to reflect the PR.