1. Introduction Problems in operations research are usually modelled as single objective ones even though there exists, in real world, several goals that should be attained. Multiple reasons of why inherently multi-objective problems are modelled as single objective ones can be outlined: better understanding of the problem features and simplification of the mathematical formulation of the problem, among others. However, summarising (usually conflicting) objectives into one objective function can be a sign of over-simplification in the problem modelling process. Multi-objective optimisation (MO) aims to optimise two or more conflicting objectives simultaneously. Unlike in single objective optimisation, where solving a problem means to find its optimal solution w.r.t. some objective function, solving a problem in a MO context means to find a set of solutions such that improving an objective without impair at least some other objective is not possible. These solutions are called efficient solutions. Mathematically, a general MO problem can be formulated as follows: min f ( x)= ( f 1 ( x ) ,f 2 ( x ) , …,f p ( x ) ) s . t . x ∈ X ⊆ℝ n (1) where x is a vector of n elements and X is the set of feasible solutions. Consequently, f ( x)= y ∈Y ⊆ℝ p is the image of solution x ∈ X in the objective space. We say that a solution x 1 ∈ X is efficient if there is no x 2 ∈ X such that f i ( x 2 )< f i ( x 1 ) for some i =1, …,p . The set of efficient solutions is denoted by X E ⊆ X . Consequently, we denote the image of ^x ∈ X E in objective space as f (^ x)= ^y , where ^y is called non- dominated point. The set of non-dominated points of the MO problem in (1) is Y N ⊆Y . In this work a model for a MO inventory location model (ILM) problem is proposed. ILM problems aim to integrate strategic decisions with tactical ones. In particular, the ILM considered in this paper seeks to integrate location decision making (strategic) with inventory policies (tactical). We consider two conflicting objectives: On the one hand, we aim to minimize the location/allocation cost, that is, the cost of installing a warehouse and, on the other hand, we want to minimize the inventory cost, that is, the cost of holding products in our warehouses and the cost of processing an order from our customers. The problem of locating/allocating customer to distribution centres is one of the most studied problems in logistics. Usually, after decision makers determine the locations to be installed, Studies in Informatics and Control, Vol. 25, No. 2, June 2016 http://www.sic.ici.ro 189 Solving a Novel Multi-Objective Inventory Location Problem by means of a Local Search Algorithm Carolina LAGOS 1 , Jorge VEGA 2 , Guillermo GUERRERO 1 , Jose-Miguel RUBIO 3 1 Pontificia Universidad Católica de Valparaíso, CHILE, {carolina.lagos.c; guillermo.guerrero.c}@mail.pucv.cl 2 Universidad de Antofagasta, CHILE, jorge.vega@uantofa.cl 3 Universidad de Playa Ancha, CHILE, jose.rubio.l@upla.cl Abstract: Problems in operations research are usually modelled as single objective ones even though there exist several goals that should be attained. Multiple reasons of why inherently multi-objective problems are modelled as single objective can be identified: better understanding of the problem features, simplification of the mathematical formulation of the problem, among others. However, summarising (usually conflicting) objectives into one objective function can be a sign of over-simplification in the problem modelling process. In this paper we extend a single objective inventory location problem formulation to a multi-objective one. We consider two conflicting objectives, namely, location cost and inventory cost. As a result, we obtain a complex multi-objective non-linear integer programming problem. Like its single objective formulation, this multi-objective problem cannot be solved by exact methods as the number of decision variables increases. Thus, a simple yet effective multi-objective local search algorithm is implemented in this paper. Keywords: MO Local Search, MO Inventory Location, Combinatorial Optimisation.