CPLEX OPL建模报错：Cannot use type dvar float+ for constraint如何解决？

Got it, let's break down why you're hitting this error and how to fix it quickly:

Why This Error Occurs #

You declared travel_time_p as a decision variable (dvar float+), but you're trying to assign it a fixed, calculated value directly in a constraint. Decision variables are meant to be values the solver optimizes during runtime—you can't treat them like static parameters by setting them equal to distance_p[p1][p2]/velocity in a constraint. CPLEX throws this error because it can't reconcile a decision variable being used as a fixed value in a constraint definition.

The Most Common (and Simple) Fix #

Since travel_time_p is just a derived value based on your known parameters (distance_p and velocity), it shouldn't be a decision variable at all. Instead, declare it as a regular parameter:

First, update the variable declaration:

float travel_time_p[plant][plant];

Then, initialize it before your constraints using an execute block in your model:

execute {
  for(var p1 in plant) {
    for(var p2 in plant) {
      if(p1 != p2) {
        travel_time_p[p1][p2] = distance_p[p1][p2]/velocity;
      }
      // Handle the case where p1 == p2 (typically set to 0, since travel time to the same plant is zero)
      else {
        travel_time_p[p1][p2] = 0;
      }
    }
  }
}

Alternatively, if you know all distance_p values upfront, you can compute travel_time_p directly in your data file instead of using an execute block.

If You Must Keep It as a Decision Variable #

In rare scenarios (e.g., if velocity is also a decision variable, or you need to account for uncertainty), you can rework the constraint to properly enforce the relationship instead of assigning a fixed value:

forall(p1 in plant, p2 in plant : p1 != p2) {
  travel_time_p[p1][p2] * velocity == distance_p[p1][p2];
}

This rearranges the equation to avoid division (which can cause numerical quirks) and sets up a valid equality constraint the solver can process, rather than trying to "hardcode" a value into a decision variable.

内容的提问来源于stack exchange，提问作者Natdanai I.