如何进一步优化ES6实现的子集和近似匹配代码？

背景 #

我有一组商品价格数据，需要找到最接近目标数值的购买组合。原本的思路是生成所有可能的购买组合（用布尔序列表示），再计算各组合的和，从中找出最接近目标值的组合。示例如下：

• 示例数据集：$10, $5, $1
• 目标值：$12
• 布尔序列组合：001, 010, 100, 011, 101, 110, 111
• 各组合和：$1, $5, $10, $6, $11, $15, $16
• 最接近的布尔序列为101，对应和为$11

问题 #

数据集较小时，用ES6实现的代码运行正常，但数据量增加后，处理时间呈指数级增长。不更换编程语言的前提下，有哪些改进方法能缩短处理时间？

当前遇到性能瓶颈的数据集：

• data：[7524.00,1.63,2000.00,23794.00,0.01,330.00,462.00,858.00,5000.00,5250.00,23162.00,23804.00,5250.00,13376.00,0.06,0.08,4000.00,10000.00,166.56,12441.00,25.00,1000.00,20000.00,2000.00,20000.00,10000.00,5468.00,20000.00];
• target：201887

当前代码 #

let possibles = [];
let sums = [];
let target = 201887;
let data = [7524.00,1.63,2000.00,23794.00,0.01,330.00,462.00,858.00,5000.00,5250.00,23162.00,23804.00];//,5250.00,13376.00,0.06,0.08,4000.00,10000.00,166.56,12441.00,25.00,1000.00,20000.00,2000.00,20000.00,10000.00,5468.00,20000.00];

let rep = new Array(data.length).fill(1).join('');

let decMax = b2d(rep) + 1;

genPossibles(decMax, possibles)
.then(genSums()
.then(matchTarget()));

/****************
    functions
****************/

function d2b(d){
    return Number(d).toString(2);
}
function b2d(b) {
    return parseInt(b, 2);
}

function genPossibles(decMax, arr, start=0, runs=1000) {
    return new Promise((resolve, reject) => {
    //console.log(decMax, arr.length, start, runs);
    for(let i = start; i < start+runs; ++i) {
      if(start+i > decMax) {
        possibles.pop(); // last item is useless
        console.log("possibles", possibles.length);
        resolve();
        return;
      }

      let binStr = d2b(i);
      possibles.push(binStr);
    }

    setTimeout(() => genPossibles(decMax, arr, start+runs), 0);
  });
}
function genSums(start=0, runs=1000) {
    return new Promise((resolve, reject) => {
    for(let i = start; i < start+runs; ++i) {
      if(i >= possibles.length) {
        console.log("sums", sums.length);
        resolve();
        return;
      }

      let str = possibles[i];
      let sum = 0;
      str.split('').reverse().forEach((bool, j) => {
        if(bool == 1) sum += +data[j];
      });
      sums.push(sum);
    }

    setTimeout(() => genSums(start+runs), 0);
  });
}
function matchTarget() {
    let closest = {id:-1, bin: '', total:-1, diff:Infinity};
    sums.forEach((o, i) => {
      let d = Math.abs(target - o);
      if(d < closest.diff) {
        closest.id = i;
        closest.bin = possibles[i];
        closest.total = o;
        closest.diff = d;
      }
    });
    console.log("closest", closest);
}

补充说明 #

目前暂时注释了部分数据，待找到优化方案后再启用完整数据集。

内容的提问来源于stack exchange，提问作者Jacksonkr