What is Operations Research?
Operations Research is a discipline at the intersection of applied mathematics and computer science. The objective of Operations Research is to find solutions (Research) to real-world problems (operational). It employs resolution methods (algorithms) to construct solutions to well-formulated problems (models).
Generally, an Operations Research problem is formulated as follows: “Find values for the unknowns x₁, x₂,…, xₙ, given that these unknowns must satisfy a set of properties (constraints) linking them together c₁, c₂, …, cₘ.” This amounts to solving a system of inequalities. In practice, the number of unknowns and inequalities can be large (> 10,000, for example), making it impossible to solve “by hand” — hence the need to use high-performance algorithms.
Optimising consists of finding a solution in the feasible domain, the best possible according to a quantitative criterion (“score”, “objective function”):
- “Find values for the unknowns x1, x2,…, xn,
- maximising a function f(x1, x2,…, xn),
- Given that these unknowns must satisfy a set of properties (constraints) linking them together, c1, c2, …, cm.”
How does Operations Research support business decision-making?
Making a decision generally requires considering a potentially large number of alternatives, each of which must be evaluated against one or more criteria to determine which one(s) are “the best”. It is to help decision-makers with these sometimes complex choices that Operations Research emerged, aiming to provide decision support tools. These models provide a representation of reality and its constraints that is easier to manipulate, a clear formulation of the criteria that guide choices, and, therefore, a rationalisation of the decisions resulting from their use.
Whether strategic and global, or operational and local, many decisions made to manage organisations aim to find the resources that will be deployed to meet needs (assignment of production tasks to industrial equipment, organisation of logistics flows, use of means to transport goods or people, mobilisation of human resources to perform services, etc.). Such decisions often involve numerous quantitative choices regarding individual components (for example, building a master production plan requires determining which product will be manufactured where, in what quantities, when, and from which site to which customer, and in what quantities and when). These choices must be consistent with one another and comply with the constraints characterising the context. As a result, each decision can be difficult to build and evaluate using traditional methods.
What are the main Operations Research approaches?
Operations Research offers a large number of algorithmic approaches for solving business problems. These approaches can be used individually or in combination, depending on the company’s needs. To support our clients, EURODECISION’s experts draw in particular on:
Linear programming and mathematical programmingHeuristics and meta-heuristicsConstraint programming
Response surface-based approaches
What is the relationship between Operations Research and Artificial Intelligence?
Operations Research (optimisation) is one of the 4 families of Artificial Intelligence, alongside statistical and connectionist AI (which includes, in particular, machine learning, symbolic AI, and collaborative AI.
Why entrust an optimisation project to EURODECISION’s experts?
For more than 30 years, our engineers and consultants have carried out numerous Operations Research projects (“pure” or based on hybrid methods) for clients across a wide range of sectors. We have developed a software offering built on highly flexible and configurable business-optimisation components to address the diverse challenges encountered.
Our company is also a founding member of ROADEF, the French society for operations research and decision support.