Polyhedral Schedulers have been widely used for loop optimization in general-purpose compilers and, more recently, deep-learning compilers. State-of-the-art scheduling algorithms define a vast space of loop transformations and find an optimal solution according to pre-designed cost functions. However, this Integer Linear Programming (ILP) approach can sometimes have complexity issues. The resolution time of ILP grows rapidly as the size of the problem increases. The complexity of the kernel to optimize, in terms of the number of statements, loops, and dependencies, has a significant impact on solving time. Additionally, the performance of ILP solvers can be severely impacted by big coefficients in the ILP, which mostly come from the loop bounds in the original kernel. To tackle this issue, this paper introduces a technique called "ParameTrick". This technique is used during the scheduling pipeline to replace some large coefficients with parameters, which are new variables in the ILP model. The approach is analyzed to determine its impact on the solving time, as well as the transformations that are lost and how to limit the number of unfeasible transformations. Results show that ParameTrick leads to faster solving times while the space of dropped desirable solutions is limited.