logo

Agendamento ponderado de tarefas | Conjunto 2 (usando LIS)

Dados N empregos onde cada trabalho é representado pelos três elementos seguintes.
1. Hora de início 
2. Hora de término 
3. Lucro ou valor associado
Encontre o subconjunto de empregos com lucro máximo, de modo que não haja sobreposição de dois empregos no subconjunto.

Exemplos:  



    Input:       
Number of Jobs n = 4
Job Details {Start Time Finish Time Profit}
Job 1: {1 2 50}
Job 2: {3 5 20}
Job 3: {6 19 100}
Job 4: {2 100 200}

Output:
Job 1: {1 2 50}
Job 4: {2 100 200}

Explanation: We can get the maximum profit by
scheduling jobs 1 and 4 and maximum profit is 250.

Em anterior post discutimos sobre o problema do agendamento ponderado de tarefas. Discutimos uma solução DP onde basicamente incluímos ou excluímos o trabalho atual. Neste post é discutida outra solução interessante de DP onde também imprimimos os Jobs. Este problema é uma variação do padrão Subsequência crescente mais longa (LIS) problema. Precisamos de uma ligeira mudança na solução de Programação Dinâmica do problema de LIS.

Primeiro precisamos classificar os trabalhos de acordo com o horário de início. Seja job[0..n-1] o array de jobs após a classificação. Definimos o vetor L tal que L[i] é ele próprio um vetor que armazena o agendamento ponderado de trabalho de trabalho[0..i] que termina com trabalho[i]. Portanto, para um índice i L[i] pode ser escrito recursivamente como - 

L[0] = {job[0]}  
L[i] = {MaxSum(L[j])} + job[i] where j < i and job[j].finish <= job[i].start
= job[i] if there is no such j


Por exemplo, considere os pares {3 10 20} {1 2 50} {6 19 100} {2 100 200}



After sorting we get   
{1 2 50} {2 100 200} {3 10 20} {6 19 100}

Therefore
L[0]: {1 2 50}
L[1]: {1 2 50} {2 100 200}
L[2]: {1 2 50} {3 10 20}
L[3]: {1 2 50} {6 19 100}

Escolhemos o vetor com maior lucro. Neste caso L[1].

Abaixo está a implementação da ideia acima – 

C++
// C++ program for weighted job scheduling using LIS #include    #include  #include    using namespace std; // A job has start time finish time and profit. struct Job {  int start finish profit; }; // Utility function to calculate sum of all vector // elements int findSum(vector<Job> arr) {  int sum = 0;  for (int i = 0; i < arr.size(); i++)  sum += arr[i].profit;  return sum; } // comparator function for sort function int compare(Job x Job y) {  return x.start < y.start; } // The main function that finds the maximum possible // profit from given array of jobs void findMaxProfit(vector<Job> &arr) {  // Sort arr[] by start time.  sort(arr.begin() arr.end() compare);  // L[i] stores Weighted Job Scheduling of  // job[0..i] that ends with job[i]  vector<vector<Job>> L(arr.size());  // L[0] is equal to arr[0]  L[0].push_back(arr[0]);  // start from index 1  for (int i = 1; i < arr.size(); i++)  {  // for every j less than i  for (int j = 0; j < i; j++)  {  // L[i] = {MaxSum(L[j])} + arr[i] where j < i  // and arr[j].finish <= arr[i].start  if ((arr[j].finish <= arr[i].start) &&  (findSum(L[j]) > findSum(L[i])))  L[i] = L[j];  }  L[i].push_back(arr[i]);  }  vector<Job> maxChain;  // find one with max profit  for (int i = 0; i < L.size(); i++)  if (findSum(L[i]) > findSum(maxChain))  maxChain = L[i];  for (int i = 0; i < maxChain.size(); i++)  cout << '(' << maxChain[i].start << ' ' <<  maxChain[i].finish << ' '  << maxChain[i].profit << ') '; } // Driver Function int main() {  Job a[] = { {3 10 20} {1 2 50} {6 19 100}  {2 100 200} };  int n = sizeof(a) / sizeof(a[0]);  vector<Job> arr(a a + n);  findMaxProfit(arr);  return 0; } 
Java
// Java program for weighted job  // scheduling using LIS import java.util.ArrayList; import java.util.Arrays; import java.util.Collections; import java.util.Comparator; class Graph{ // A job has start time finish time // and profit. static class Job {  int start finish profit;  public Job(int start int finish   int profit)  {  this.start = start;  this.finish = finish;  this.profit = profit;  } }; // Utility function to calculate sum of all // ArrayList elements static int findSum(ArrayList<Job> arr)  {  int sum = 0;    for(int i = 0; i < arr.size(); i++)  sum += arr.get(i).profit;    return sum; } // The main function that finds the maximum // possible profit from given array of jobs static void findMaxProfit(ArrayList<Job> arr) {    // Sort arr[] by start time.  Collections.sort(arr new Comparator<Job>()   {  @Override  public int compare(Job x Job y)   {  return x.start - y.start;  }  });    // sort(arr.begin() arr.end() compare);  // L[i] stores Weighted Job Scheduling of  // job[0..i] that ends with job[i]  ArrayList<ArrayList<Job>> L = new ArrayList<>();  for(int i = 0; i < arr.size(); i++)  {  L.add(new ArrayList<>());  }  // L[0] is equal to arr[0]  L.get(0).add(arr.get(0));  // Start from index 1  for(int i = 1; i < arr.size(); i++)   {    // For every j less than i  for(int j = 0; j < i; j++)  {    // L[i] = {MaxSum(L[j])} + arr[i] where j < i  // and arr[j].finish <= arr[i].start  if ((arr.get(j).finish <= arr.get(i).start) &&  (findSum(L.get(j)) > findSum(L.get(i))))  {  ArrayList<Job> copied = new ArrayList<>(  L.get(j));  L.set(i copied);  }  }  L.get(i).add(arr.get(i));  }  ArrayList<Job> maxChain = new ArrayList<>();  // Find one with max profit  for(int i = 0; i < L.size(); i++)  if (findSum(L.get(i)) > findSum(maxChain))  maxChain = L.get(i);  for(int i = 0; i < maxChain.size(); i++)   {  System.out.printf('(%d %d %d)n'   maxChain.get(i).start   maxChain.get(i).finish  maxChain.get(i).profit);  } } // Driver code public static void main(String[] args) {  Job[] a = { new Job(3 10 20)   new Job(1 2 50)  new Job(6 19 100)  new Job(2 100 200) };  ArrayList<Job> arr = new ArrayList<>(  Arrays.asList(a));  findMaxProfit(arr); } } // This code is contributed by sanjeev2552 
Python
# Python program for weighted job scheduling using LIS import sys # A job has start time finish time and profit. class Job: def __init__(self start finish profit): self.start = start self.finish = finish self.profit = profit # Utility function to calculate sum of all vector elements def findSum(arr): sum = 0 for i in range(len(arr)): sum += arr[i].profit return sum # comparator function for sort function def compare(x y): if x.start < y.start: return -1 elif x.start == y.start: return 0 else: return 1 # The main function that finds the maximum possible profit from given array of jobs def findMaxProfit(arr): # Sort arr[] by start time. arr.sort(key=lambda x: x.start) # L[i] stores Weighted Job Scheduling of job[0..i] that ends with job[i] L = [[] for _ in range(len(arr))] # L[0] is equal to arr[0] L[0].append(arr[0]) # start from index 1 for i in range(1 len(arr)): # for every j less than i for j in range(i): # L[i] = {MaxSum(L[j])} + arr[i] where j < i # and arr[j].finish <= arr[i].start if arr[j].finish <= arr[i].start and findSum(L[j]) > findSum(L[i]): L[i] = L[j][:] L[i].append(arr[i]) maxChain = [] # find one with max profit for i in range(len(L)): if findSum(L[i]) > findSum(maxChain): maxChain = L[i] for i in range(len(maxChain)): print('({} {} {})'.format( maxChain[i].start maxChain[i].finish maxChain[i].profit) end=' ') # Driver Function if __name__ == '__main__': a = [Job(3 10 20) Job(1 2 50) Job(6 19 100) Job(2 100 200)] findMaxProfit(a) 
C#
using System; using System.Collections.Generic; using System.Linq; public class Graph {  // A job has start time finish time  // and profit.  public class Job  {  public int start finish profit;  public Job(int start int finish   int profit)  {  this.start = start;  this.finish = finish;  this.profit = profit;  }  };  // Utility function to calculate sum of all  // ArrayList elements  public static int FindSum(List<Job> arr)   {  int sum = 0;    for(int i = 0; i < arr.Count; i++)  sum += arr.ElementAt(i).profit;    return sum;  }  // The main function that finds the maximum  // possible profit from given array of jobs  public static void FindMaxProfit(List<Job> arr)  {    // Sort arr[] by start time.  arr.Sort((x y) => x.start.CompareTo(y.start));  // L[i] stores Weighted Job Scheduling of  // job[0..i] that ends with job[i]  List<List<Job>> L = new List<List<Job>>();  for(int i = 0; i < arr.Count; i++)  {  L.Add(new List<Job>());  }  // L[0] is equal to arr[0]  L[0].Add(arr[0]);  // Start from index 1  for(int i = 1; i < arr.Count; i++)   {    // For every j less than i  for(int j = 0; j < i; j++)  {    // L[i] = {MaxSum(L[j])} + arr[i] where j < i  // and arr[j].finish <= arr[i].start  if ((arr[j].finish <= arr[i].start) &&  (FindSum(L[j]) > FindSum(L[i])))  {  List<Job> copied = new List<Job>(  L[j]);  L[i] = copied;  }  }  L[i].Add(arr[i]);  }  List<Job> maxChain = new List<Job>();  // Find one with max profit  for(int i = 0; i < L.Count; i++)  if (FindSum(L[i]) > FindSum(maxChain))  maxChain = L[i];  for(int i = 0; i < maxChain.Count; i++)   {  Console.WriteLine('({0} {1} {2})'   maxChain[i].start   maxChain[i].finish  maxChain[i].profit);  }  }  // Driver code  public static void Main(String[] args)  {  Job[] a = { new Job(3 10 20)   new Job(1 2 50)  new Job(6 19 100)  new Job(2 100 200) };  List<Job> arr = new List<Job>(a);  FindMaxProfit(arr);  } } 
JavaScript
// JavaScript program for weighted job scheduling using LIS // A job has start time finish time and profit. function Job(start finish profit) {  this.start = start;  this.finish = finish;  this.profit = profit; } // Utility function to calculate sum of all vector // elements function findSum(arr) {  let sum = 0;  for (let i = 0; i < arr.length; i++) {  sum += arr[i].profit;  }  return sum; } // comparator function for sort function function compare(x y) {  return x.start < y.start; } // The main function that finds the maximum possible // profit from given array of jobs function findMaxProfit(arr) {  // Sort arr[] by start time.  arr.sort(compare);  // L[i] stores Weighted Job Scheduling of  // job[0..i] that ends with job[i]  let L = new Array(arr.length).fill([]);  // L[0] is equal to arr[0]  L[0] = [arr[0]];  // start from index 1  for (let i = 1; i < arr.length; i++) {  // for every j less than i  for (let j = 0; j < i; j++) {  // L[i] = {MaxSum(L[j])} + arr[i] where j < i  // and arr[j].finish <= arr[i].start  if (arr[j].finish <= arr[i].start && findSum(L[j]) > findSum(L[i])) {  L[i] = L[j];  }  }  L[i].push(arr[i]);  }  let maxChain = [];  // find one with max profit  for (let i = 0; i < L.length; i++) {  if (findSum(L[i]) > findSum(maxChain)) {  maxChain = L[i];  }  }  for (let i = 0; i < maxChain.length; i++) {  console.log(  '(' +  maxChain[i].start +  ' ' +  maxChain[i].finish +  ' ' +  maxChain[i].profit +  ') '  );  } } // Driver Function let a = [  new Job(3 10 20)  new Job(1 2 50)  new Job(2 100 200) ]; findMaxProfit(a); 

Saída
(1 2 50) (2 100 200) 


Podemos otimizar ainda mais a solução DP acima removendo a função findSum(). Em vez disso, podemos manter outro vetor/array para armazenar a soma do lucro máximo possível até o trabalho i.



Complexidade de tempo da solução de programação dinâmica acima é O (n2) onde n é o número de trabalhos. 
Espaço auxiliar usado pelo programa é O(n2).