In this context, "optimize" is to select (find) the "best" solution among many possible alternatives. This optimization process can be viewed as a sequence of decisions that provide the right solution. If given a sub-making,
always knows what is the decision to be made then for the optimal sequence, the problem is trivially resolved elemental and take a decision after another, which is known as greedy strategy. <= pesos.length ; j++)
<= capacidad; c++){
often but is not possible to implement the greedy strategy, it meets the Bellman optimality principle: "given an optimal sequence of decisions, any subsequence of it is, in turn, optimal." In this case it is still possible to go by taking basic decisions in < pesos[j-1] ){
confident that the combination of them will remain optimal, but will then need to explore many sequences of decisions to make with the right, being here where dynamic programming.
Consider a problem as a sequence of decisions is equivalent to divide into sub undersized, in principle, more easily solvable. This is part of the Divide and Conquer method, technique similar to that of Dynamic Programming. Dynamic programming is applied when the subdivision of a problem leads to:
A huge number of subproblems.
subproblems whose solutions partial overlap. Groups
subproblems very different complexity.
circumstances which together or separately, would lead to an exponential complexity of the strategy "Divide and Conquer."
0 comments:
Post a Comment