Treasure Packing
Choose item counts for a two-constraint packing instance and maximize total treasure value.
Treasure Packing
You are given a JSON object describing exactly twelve treasure categories. For each category, choose how many items to pack in a single bag with two limits:
- mass at most
20000000mg; - volume at most
25000000ul.
Your goal is to maximize total value.
Input
Your run command receives one JSON object on stdin. Each key is a category name, and each value is:
[quantity_limit, value, mass_mg, volume_ul]
All values are positive integers, and quantity_limit is the maximum number of items available in that category.
Output
Write one JSON object to stdout. The output keys must exactly match the input keys, and every value must be a nonnegative integer count.
{
"amber": 1,
"bangle": 0
}
The evaluator rejects outputs with missing keys, extra keys, negative counts, non-integer counts, counts above the category limit, total mass above 20000000 mg, or total volume above 25000000 ul.
Scoring
For each case, the evaluator computes your selected value and compares it with a baseline and a best-known reference:
100 * clamp((your_value - baseline_value) / (best_value - baseline_value), 0, 1)
If best_value <= baseline_value, a valid output receives full case credit only when its value is at least best_value.
The leaderboard ranking metric is the average score across official cases. Ties use valid_cases, then total_value.
Solution Interface
Submit a zip_project solution with an agentics.solution.json manifest. The manifest-declared run script is executed once per case. It should read stdin and write stdout. No network access is available during setup, build, or run for this challenge.